Tasks/Image generation

Choose by task

Image generation tools compared

Read the assessment, then compare prices, capabilities and constraints.

This topic covers 28 published image generation assessments: 3 support guest export without login or watermarks, 9 deliver watermark-free files, and 1 offer sourced commercial-use plans. All scores are bound to fixed benchmark fixtures and dated native exports.

Quick scenarios:

28 published assessments

Plus 5 tested tools withheld from ranking due to free-tier paywalls (B1) or backend errors (B2) View blocked records ↓

Test scope & pricing dates

Pricing shows its own verification date and scope; tests retain their original dates. Vendor claims and test results are separate; paid specifications may be untested. About the scope →

magnific.ai

magnific.ai · Image generation

95/100

Seven-aspect task score

Magnific AI (by Freepik) delivers outstanding multi-constraint reasoning and precise typographic rendering. Given a demanding prompt requiring exact English spelling, count-color binding, and a 5-fingered brass robot hand, the engine accurately rendered the carved 'REALTEST 2026' wooden sign with 3 red apples and 2 blue mugs. The server wait was 3.85 seconds, offering a clean 1024x1024 unwatermarked PNG deliverable.

Tested export 1024 × 1024 · PNG

Historical test · 2026-11-02
Account / tier unconfirmed · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (1024 × 1024 · PNG) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Flawless text spelling on the wooden sign (25/25); exact object counting with zero color bleeding (20/20); highly coherent 5-finger mechanical hand anatomy (20/20).

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Flawless text spelling on the wooden sign (25/25); exact object counting with zero color bleeding (20/20); highly coherent 5-finger mechanical hand anatomy (20/20).

Original historical observation

Not verified

Suitable uses · historical assessment

Not verified

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

1024 × 1024 · PNG · Check watermark record

Detailed facts & historical record

Recorded export specifications: 1024 × 1024 · PNG · Check watermark record
Historical specification note · not an official maximum: 1024x1024 (1K watermark-free PNG)
Watermark observation: Not verified
Batch & API: Not verified
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: 1024 × 1024 · PNG

Evidence & scope →

Speed & reliability

End-to-end time unrecorded; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 3.85s
Date & fixture: 2026-11-02 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Account requirement unconfirmed

Detailed facts & historical record

Account requirements: Not verified
Sign-in methods: Not verified
Manual adjustment: Not verified
Interaction observations: Not verified

Evidence & scope →

Background & reputation

Freepik Company S.L. (Magnific AI) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Freepik Company S.L. (Magnific AI)
Recorded origin: Not verified
Domain history record: Not verified
Background & reputation assessment: Not verified
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Processing location unconfirmed；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Processing location unconfirmed
Third-party tracking record: Google Tag Manager · Stripe · Segment
Engineering & privacy assessment: Not verified
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Magnific web credit subscription and generative upscaling service · 2026-11-01

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free Tier2026-11-01 ↗

Free

Free Daily basic AI image generations and light trials

Basic image generation and community asset browsing；Excludes advanced high-res upscaling credits; personal non-commercial use only；Permanently free, no renewal；Perpetual access

Premium Monthly2026-11-01 ↗

USD 20/month

—

Monthly 20,000 credits per month for upscaling, image, video, and audio models

Access to image/video/audio models, shared canvas, commercial license, and pro editing；Credits reset monthly without rollover；Auto-renews monthly; cancel anytime

Rules [1]

Premium Annual2026-11-01 ↗

—

Annual price not verified

Annual 240,000 credits per year valid for 1 full year with no monthly resets

Full Premium features with 25% annual discount equivalent；Displays monthly equivalent of $14.50/mo; exact lump-sum annual charge not printed on card

Rules [2] [3]

Premium+ Monthly2026-11-01 ↗

USD 45/month

—

Monthly 45,000 credits per month with unlimited base generations (Nano Banana 2 1K)

Full model access, unlimited base generation, commercial AI + music rights；Credits reset monthly

Rules [1] [4]

Premium+ Annual2026-11-01 ↗

—

Annual price not verified

Annual 600,000 credits per year valid for 1 year with no monthly resets

Full Premium+ features with annual discount equivalent；Displays monthly equivalent of $33.75/mo; exact annual charge not printed on card

Rules [2] [3]

Pro Starter Monthly2026-11-01 ↗

USD 110/month

—

Monthly 112,500 credits per month (upgradable to 300,000), unused credits roll over 1 month

Pro heavy upscaling, credit rollover, unlimited 2K generation, and pro editing；Unused credits roll over until end of next billing month

Rules [1] [4]

Pro Starter Annual2026-11-01 ↗

—

Annual price not verified

Annual 1,500,000 credits per year valid for 1 full year

Full Pro Starter features with annual discount；Displays monthly equivalent of $82.50/mo; exact annual charge not printed on card

Rules [2] [3]

7 plan / billing-period records, all listed; — means that period is not recorded.

[1] Active during billing month

[2] Auto-renews annually

[3] Active during subscription year

[4] Auto-renews monthly

Offers monthly and annual subscriptions; monthly plans include recurring monthly credits while annual plans issue lump-sum yearly credits with a 25% discount (displayed as monthly equivalents).

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Cancellation and refund terms unverified

Magnific Official Pricing Page (Annual) ↗ · 2026-11-01 Capture 1 ↗

Magnific Official Pricing Page (Monthly) ↗ · 2026-11-01 Capture 1 ↗

Allowance & disclosure on test date

Not recorded

Evaluation basis & full scores · Overall 95.1/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 86/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	32 / 35	Explanation unavailable	Evidence A →
Interaction restraint	24 / 30	Explanation unavailable	Evidence A →
Background & reputation	18 / 20	Explanation unavailable	Evidence B →
Engineering & privacy	12 / 15	Explanation unavailable	Evidence A →

Capability · 99/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	65 / 65	Explanation unavailable	Evidence A →
Deliverable specifications	20 / 20	Forced watermark in free export (0/8); Export & Max deliverable spec: 1024x1024 (1K watermark-free PNG)	Evidence A →
Processing speed	14 / 15	Network-observed server wait: 3.85s (API mode: async_verified; end-to-end N/A shown for reference only, never penalized)	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

burningbridgesai.com

burningbridgesai.com · Image generation

94/100

Seven-aspect task score

Burning Bridges AI offers 18 trial credits (3 images); its GPT Image 2.5 Sunburst delivers an unwatermarked 1254×1254 PNG in 5.67s server time, achieving flawless prompt adherence across 'REALTEST 2026' text typography, multi-object counts, and 5-finger robot hand anatomy.

Tested export 1254 × 1254 · PNG

Historical test · 2026-11-01
Signed-in flow; paid tier not standardized · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

burningbridgesai.com native output thumbnail

Inspect native image (1254 × 1254 · PNG) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Flawless multi-constraint execution: engraved 'REALTEST 2026' wooden sign typography, precise 3 red apples and 2 blue ceramic mugs binding, 5-finger articulated brass robot hand, and clean 1254px PNG delivery.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Original historical observation

Burning Bridges AI delivers stellar performance on anchor_t2i_v1: 5.67s cloud inference generates an unwatermarked 1254×1254 PNG with flawless 'REALTEST 2026' text, precise counts of 3 red apples and 2 blue ceramic mugs, and a 5-fingered brass robot hand.

Suitable uses · historical assessment

Creators requiring reliable multi-object prompt adherence, accurate typography rendering, and clean PNG downloads

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

1254 × 1254 · PNG · Check watermark record

Detailed facts & historical record

Recorded export specifications: 1254 × 1254 · PNG · Check watermark record
Historical specification note · not an official maximum: 1254×1254 (PNG)
Watermark observation: Completely free of visible watermarks
Batch & API: Supports My creations history drawer, parameter review, and native full-size preview downloads
Recorded free-tier dimensions: 1254 × 1254
Attached file dimensions: 1254 × 1254 · PNG

Evidence & scope →

Speed & reliability

About 6.2s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 5.67s · End-to-end reference: 6.2s
Date & fixture: 2026-11-01 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Supports Google SSO and email login
Manual adjustment: Not verified
Interaction observations: Studio workspace offers clear model selection (GPT Image 2.5 Sunburst), creation modes, prompt input, and settings, estimating 6 credits upfront with an integrated My creations drawer and native Download button.

Evidence & scope →

Background & reputation

Burning Bridges AI · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Burning Bridges AI
Recorded origin: Not verified
Domain history record: Not verified
Background & reputation assessment: Inherited from Burning Bridges AI product baseline: Emerging trend-focused web platform with clear terms, privacy disclosures, and a 72-hour refund policy, though formal enterprise entity data is minimal.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Google Tag Manager · Stripe
Engineering & privacy assessment: Server inference wait measured at 5.67s (POST /api/generate-image -> polling GET /api/generate-image/status?id=oZuE4FHxdUGHsGnov2VgF), exporting an unwatermarked 1254×1254 (2.72MB) PNG with zero third-party tracking on delivery.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Burning Bridges AI credit subscriptions and one-time pack pricing · 2026-11-01

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Basic Tier2026-11-01 ↗

Free

Free 18 one-time signup credits (up to 3 images or 1 video)

Sample video and image generation with history included；Watermarked previews and downloads; clean downloads require paid plan；One-time trial, no auto-renewal；Valid until trial credits consumed

Starter Monthly2026-11-01 ↗

USD 29.9/month

—

Monthly 800 credits/month (up to 133 images or 44 videos)

Watermark-free downloads, Seedance 2.5 and Kling 3.0, up to 4K image and HD video

Rules [1] [2] [3]

Pro Monthly2026-11-01 ↗

USD 49.9/month

—

Monthly 1,800 credits/month (up to 300 images or 100 videos)

Watermark-free downloads, generous credit allotment across models

Rules [1] [2] [3]

Starter Annual2026-11-01 ↗

—

USD 100/year

Annual 2,400 credits/year ($8.3/mo equivalent, up to 400 images or 133 videos)

Deep annual discount, watermark-free downloads, 4K image and HD video creation；Billed annually at $100.0 upfront；Renews annually；Valid within annual billing period

4 plan / billing-period records, all listed; — means that period is not recorded.

[1] Subscription credits reset monthly

[2] Renews monthly, cancel anytime in billing settings

[3] Valid within monthly billing period

Offers monthly/annual subscriptions and non-expiring one-time packs. New signups receive 18 trial credits (watermarked). Annual plans bill upfront with up to 50% discounts.

Cancellation, sources & verification record

Cancellation and refund terms unverified

Burning Bridges AI Pricing Page (Annual) ↗ · 2026-11-01 Capture 1 ↗

Burning Bridges AI Pricing Page (Monthly) ↗ · 2026-11-01 Capture 1 ↗

Allowance & disclosure on test date

18 trial credits on signup, 6 credits per image (up to 3 free generations)

Estimated credits displayed prominently as '6 credits' before generating

Evaluation basis & full scores · Overall 94.3/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 88/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	33 / 35	Inherited from Burning Bridges AI product baseline: Fully transparent public pricing across monthly, annual (with upfront totals), and one-time credit packs with clear per-generation costs.	Evidence A →
Interaction restraint	28 / 30	Studio workspace offers clear model selection (GPT Image 2.5 Sunburst), creation modes, prompt input, and settings, estimating 6 credits upfront with an integrated My creations drawer and native Download button.	Evidence A →
Background & reputation	13 / 20	Inherited from Burning Bridges AI product baseline: Emerging trend-focused web platform with clear terms, privacy disclosures, and a 72-hour refund policy, though formal enterprise entity data is minimal.	Evidence A →
Engineering & privacy	14 / 15	Server inference wait measured at 5.67s (POST /api/generate-image -> polling GET /api/generate-image/status?id=oZuE4FHxdUGHsGnov2VgF), exporting an unwatermarked 1254×1254 (2.72MB) PNG with zero third-party tracking on delivery.	Evidence A →

Capability · 97/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	65 / 65	Text rendering (25.0/25: Sign precisely carves 'REALTEST 2026' across two lines with natural wood integration and zero spelling errors) · Count & color binding (20.0/20: Exactly 3 red apples and 2 blue ceramic mugs on oak table without bleeding) · Structure & composition (20.0/20: 5 distinct mechanical joints on the vintage brass robot right hand holding the wooden sign).	Evidence A →
Deliverable specifications	20 / 20	Clean unwatermarked export (8/8); native deliverable reaches 1254×1254 high-resolution PNG (12/12).	Evidence A →
Processing speed	12 / 15	Verified server processing time 5.67s (good 4.0–8.0s, async_verified, 6.20s end-to-end) (12.0/15).	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Recraft

recraft.ai · Image generation

94/100

Seven-aspect task score

Text, counts and composition perform well, with prompts preserved through login into canvas editing; an account and credits are needed for this 1024px clean export.

Tested export 1024×1024 (1.62MB unwatermarked PNG, 100% pixel retention) · Historical description; output not attached

Historical test · 2026-10-16
Signed-in flow; paid tier not standardized · Fixed English prompt · Historical page record; output not attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Entry & rules ↗Result screen ↗Download & limits ↗Additional screenshot 1 ↗

Use this capability ↗Tool profile →Full report →

Task quality

Text, counts and hand constraints matched this prompt; repeatability still needs multiple runs.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Text, counts and hand constraints matched this prompt; repeatability still needs multiple runs.

Original historical observation

Achieved a perfect 65.0/65 constraint score (#1 overall at 94.1 pts): not only did Recraft render exact "REALTEST 2026" text, 3 red apples + 2 blue mugs, and a 5-digit brass hand, it even rendered an upside-down reflection of the text on the oak table.

Suitable uses · historical assessment

Designers and creators needing #1 SOTA prompt adherence (65.0/65), flawless typography with physical reflections, vector/SVG workflows, and unwatermarked 1024×1024 PNG exports

Result & output evidence

No native output obtained; the preview does not establish final export quality.

Open result screenshot ↗

Full test record →

Evidence & scope →

Delivery capabilities

1024×1024 (1.62MB unwatermarked PNG, 100% pixel retention) · Historical description; output not attached · Check watermark record

Detailed facts & historical record

Recorded export specifications

1024×1024 (1.62MB unwatermarked PNG, 100% pixel retention) · Historical description; output not attached · Check watermark record

No native output is attached; check the evidence for specification and watermark claims.

Historical specification note · not an official maximum

1024×1024 (Unwatermarked PNG · Free 60 Credits on Sign-in)

Watermark observation

Zero watermark (clean 1024×1024 PNG export)

Batch & API

Infinite canvas with native Vector/SVG vectorizer, Mockup generator, Background Remover, Eraser, and REST API

Recorded free-tier dimensions

Not verified

Attached file dimensions

Not verified

Evidence & scope →

Speed & reliability

About 8.65s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 8.2s · End-to-end reference: 8.65s
Date & fixture: 2026-10-16 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Free Google / Discord / Apple / Email login (grants 60 initial free credits)
Manual adjustment: Not verified
Interaction observations: Zero deceptive paywalls or forced watermarks: after 1-click Google login, the prompt is automatically populated on the infinite canvas and generates a 1024×1024 image in ~8.2s (costing 4 credits, leaving 56 credits).

Evidence & scope →

Background & reputation

Recraft, Inc. · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Recraft, Inc.
Recorded origin: London, United Kingdom
Domain history record: 2022–present
Background & reputation assessment: Operated by Recraft, Inc. (London, UK, founded 2022 by Anna Veronika Dorogush, co-creator of CatBoost), creator of the SOTA Recraft V3/V4 foundation models for graphic design and typography.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Exports a lossless 1024×1024 (1.62MB) unwatermarked PNG directly from the infinite canvas in one click.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

All web Basic/Pro/Teams volumes; Teams per seat; API excluded · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Basic2026-10-20 ↗

USD 12/month

USD 120/year

Monthly 1,000 general credits/month; Teams billed per seat Rules [1]

Annual 1,000 general credits/month; Teams billed per seat Rules [1]

Pro 2k2026-10-20 ↗

USD 20/month

USD 192/year

Monthly 2,000 general credits/month; Teams billed per seat Rules [1]

Annual 2,000 general credits/month; Teams billed per seat Rules [1]

Teams 2k / seat2026-10-20 ↗

USD 22/month

USD 211/year

Monthly 2,000 general credits/month; Teams billed per seat Rules [1]

Annual 2,000 general credits/month; Teams billed per seat Rules [1]

Pro 4k2026-10-20 ↗

USD 40/month

USD 384/year

Monthly 4,000 general credits/month Rules [2] [3]

Annual 4,000 general credits/month Rules [2] [3]

Teams 4k2026-10-20 ↗

USD 44/month

USD 422/year

Monthly 4,000 general credits/month per seat Rules [3] [4]

Annual 4,000 general credits/month per seat Rules [3] [4]

Pro 8k2026-10-20 ↗

USD 80/month

USD 768/year

Monthly 8,000 general credits/month Rules [2] [3]

Annual 8,000 general credits/month Rules [2] [3]

Teams 8k2026-10-20 ↗

USD 88/month

USD 844/year

Monthly 8,000 general credits/month per seat Rules [3] [4]

Annual 8,000 general credits/month per seat Rules [3] [4]

Pro 16k2026-10-20 ↗

USD 160/month

USD 1,536/year

Monthly 16,000 general credits/month Rules [2] [3]

Annual 16,000 general credits/month Rules [2] [3]

Teams 16k2026-10-20 ↗

USD 176/month

USD 1,688/year

Monthly 16,000 general credits/month per seat Rules [3] [4]

Annual 16,000 general credits/month per seat Rules [3] [4]

18 plan / billing-period records, all listed; — means that period is not recorded.

[1] Subscription credits do not roll over; purchased top-ups do not expire

[2] Annual total or monthly charge directly shown by official pricing; VAT excluded

[3] Reset each billing cycle; no rollover

[4] Annual total or monthly charge directly shown by official pricing; VAT excluded; per seat

All 2k/4k/8k/16k Pro/Teams monthly and annual tiers are listed. Top-ups require Basic/Pro/Teams, but the public page does not quote pack prices or quantify free removal counts/per-removal cost. Subscription credits do not roll over; purchased top-ups never expire. Free-plan generated images are public and not licensed for commercial use; this statement is not generalized to rights in uploaded material.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Cancel anytime, effective at period end. Discretionary refunds may be requested within 30 days with no use since payment or under 30 lifetime credits used.

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗

All credit volumes and usage rules ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗ · Capture 4 ↗ · Capture 5 ↗ · Capture 6 ↗ · Capture 7 ↗ · Capture 8 ↗

Allowance & disclosure on test date

60 free credits upon registration; each 1024×1024 generation costs 4 credits (15 free full-res images initially) plus daily credit refreshes

Prompts for free login upon clicking "Generate for free" and clearly displays remaining credit balance (56) and per-run cost (4) inside the workspace

Evaluation basis & full scores · Overall 94.1/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 92/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	30 / 35	Landing page (01_landing.png) provides an interactive prompt box and aspect ratio selector with a "Generate for free" CTA; clicking it redirects to id.recraft.ai for free Google sign-in (01_pass1_login_modal.png), which immediately grants 60 free credits and carries the typed prompt directly into the infinite canvas.	Evidence A →
Interaction restraint	28 / 30	Zero deceptive paywalls or forced watermarks: after 1-click Google login, the prompt is automatically populated on the infinite canvas and generates a 1024×1024 image in ~8.2s (costing 4 credits, leaving 56 credits).	Evidence A →
Background & reputation	19 / 20	Operated by Recraft, Inc. (London, UK, founded 2022 by Anna Veronika Dorogush, co-creator of CatBoost), creator of the SOTA Recraft V3/V4 foundation models for graphic design and typography.	Evidence C →
Engineering & privacy	15 / 15	Exports a lossless 1024×1024 (1.62MB) unwatermarked PNG directly from the infinite canvas in one click.	Evidence A →

Capability · 95/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	65 / 65	Text Rendering (25/25: Flawless 2-line "REALTEST 2026" wood-carved typography—and remarkably, the polished dark oak table beneath the sign renders a physically accurate upside-down specular reflection of "REALTEST 2026" (25/25)) · Count & Color Binding (20/20: Strictly 3 glossy red apples on the left + 2 speckled blue ceramic mugs on the right on a dark oak table, zero color bleeding (20/20)) · 5-Finger Hand & Structure (20/20: Anatomically complete vintage brass robotic right hand showing 1 thumb + 4 articulated brass fingers (5 distinct digits) gripping the wooden sign, plus realistic tungsten-bulb eyes and brass rivets (20/20))	Evidence B →
Deliverable specifications	20 / 20	Clean unwatermarked export (8/8); Export & Max deliverable spec: 1024×1024 (Unwatermarked PNG · Free 60 Credits on Sign-in)	Evidence B →
Processing speed	10 / 15	Network-observed server wait: 8.2s (API mode: async_poll; end-to-end 8.65s shown for reference only, never penalized)	Evidence B →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Krea

krea.ai · Image generation

93/100

Seven-aspect task score

Free daily credits cover Krea 3 Fast (~36 credits) to generate and export a native 2048×2048 (2K) unwatermarked PNG in 10.49s server wait with accurate text, object counts, and 5-digit hand anatomy; sign-in required.

Tested export 2048 × 2048 · PNG

Historical test · 2026-10-21
Signed-in flow; paid tier not standardized · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (2048 × 2048 · PNG) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

'REALTEST 2026' engraved text, 3 red apples + 2 blue ceramic mugs, and 5-digit brass mechanical right hand are all accurately rendered (64.5/65).

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

'REALTEST 2026' engraved text, 3 red apples + 2 blue ceramic mugs, and 5-digit brass mechanical right hand are all accurately rendered (64.5/65).

Original historical observation

Top-tier multi-constraint fulfillment (64.5/65): 'REALTEST 2026' lettering, 3 red apples + 2 blue ceramic mugs, and 5-digit brass right hand are all rendered accurately at native 2048×2048 resolution.

Suitable uses · historical assessment

Designers and creators needing 2048×2048 (2K) native PNG output with near-flawless typography, object counting, and mechanical hand anatomy

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

2048 × 2048 · PNG · Check watermark record

Detailed facts & historical record

Recorded export specifications: 2048 × 2048 · PNG · Check watermark record
Historical specification note · not an official maximum: 2048×2048 (2K PNG)
Watermark observation: No watermark (native 2048×2048 RGB PNG export)
Batch & API: Supports 1–4 batch size, LoRA & style transfer attachments, and one-click transfer to Enhancer, Edit, or Video
Recorded free-tier dimensions: 2048 × 2048
Attached file dimensions: 2048 × 2048 · PNG

Evidence & scope →

Speed & reliability

About 10.49s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 10.49s · End-to-end reference: 10.49s
Date & fixture: 2026-10-21 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Supports free Google or Email sign-in on /image
Manual adjustment: Not verified
Interaction observations: Clean dark-mode studio interface with direct controls for aspect ratio (Auto), thinking effort (Fast), and model (Krea 3); hovering the generated image reveals a one-click Download button without paywall modals.

Evidence & scope →

Background & reputation

Krea AI Inc. (San Francisco, CA) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Krea AI Inc. (San Francisco, CA)
Recorded origin: San Francisco, CA, USA
Domain history record: 2022–present
Background & reputation assessment: Developed by Krea AI Inc. (San Francisco, CA, backed by a16z), a premier Silicon Valley generative visual studio.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Dispatches async generation via /api/jobs/v2/new/k3 (network-observed server wait 10.49s) and delivers a native 2048×2048 RGB PNG with exceptional multi-constraint adherence.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Individual web plans and default Business 80k volume; excludes API and other business volumes · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free2026-10-20 ↗

Free

Free 100 general compute units/day

Basic2026-10-20 ↗

USD 9/month

USD 63/year

Monthly 5,000 general compute units/month; not removal counts

Annual 5,000 general compute units/month

Observed first-year promotion USD 63; then USD 84/year until cancelled

Rules [1]

Pro2026-10-20 ↗

USD 35/month

USD 252/year

Monthly 20,000 general compute units/month; not removal counts

Annual 20,000 general compute units/month

Observed first-year promotion USD 252; then USD 336/year until cancelled

Rules [1]

Max 60k2026-10-20 ↗

USD 105/month

USD 756/year

Monthly 60,000 general compute units/month; not removal counts

Annual 60,000 general compute units/month

Observed first-year promotion USD 756; then USD 1008/year until cancelled

Rules [1]

Business 80k2026-10-20 ↗

USD 200/month

USD 1,920/year

Monthly 80,000 general compute units/month; not removal counts

Annual 80,000 general compute units/month

Max 40k2026-10-20 ↗

USD 70/month

USD 504/year

Annual 40,000 general compute units/month

Observed first-year promotion USD 504; then USD 672/year until cancelled

Rules [1]

Monthly 40,000 general compute units/month

Max 80k2026-10-20 ↗

USD 135/month

USD 972/year

Annual 80,000 general compute units/month

Observed first-year promotion USD 972; then USD 1296/year until cancelled

Rules [1]

Monthly 80,000 general compute units/month

Max 100k2026-10-20 ↗

USD 165/month

USD 1,188/year

Annual 100,000 general compute units/month

Observed first-year promotion USD 1188; then USD 1584/year until cancelled

Rules [1]

Monthly 100,000 general compute units/month

Enterprise2026-10-20 ↗

Contact sales

Quoted

Quoted Custom enterprise support, security and administration

16 plan / billing-period records, all listed; — means that period is not recorded.

[1] First-term offer observed at checkout; eligibility, timing, currency and final tax must be checked before payment

All 10 previously missing amounts are confirmed in USD checkout summaries. Individual annual offers separate first-year and renewal charges; Business 80k is $1,920/year. Max monthly purchase paths show additional annual offers whose checkout totals were not independently confirmed; these are not merged with the default annual offer. Per-removal compute cost and extra top-up pricing are not separately specified in the verified plans/checkouts. Other Business volumes are outside scope.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Self-serve cancellation in account settings

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗

USD checkout summaries reached from official pricing ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗ · Capture 4 ↗ · Capture 5 ↗ · Capture 6 ↗ · Capture 7 ↗ · Capture 8 ↗ · Capture 9 ↗ · Capture 10 ↗

Additional annual offers on monthly purchase paths ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗

Allowance & disclosure on test date

Free account daily compute allowance covers Krea 3 (Fast, ~36 credits/image) with direct 2048×2048 PNG export

Generate button tooltip explicitly discloses '~36 credits' before submission

Evaluation basis & full scores · Overall 93/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 87/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	28 / 35	Requires a free login to access /image; hovering the generate button transparently displays '~36 credits' for Krea 3 (Fast, 2K), and the free daily credit allowance covers generating and directly exporting a 2048×2048 unwatermarked PNG (8.16MB).	Evidence A →
Interaction restraint	27 / 30	Clean dark-mode studio interface with direct controls for aspect ratio (Auto), thinking effort (Fast), and model (Krea 3); hovering the generated image reveals a one-click Download button without paywall modals.	Evidence A →
Background & reputation	18 / 20	Developed by Krea AI Inc. (San Francisco, CA, backed by a16z), a premier Silicon Valley generative visual studio.	Evidence B →
Engineering & privacy	14 / 15	Dispatches async generation via /api/jobs/v2/new/k3 (network-observed server wait 10.49s) and delivers a native 2048×2048 RGB PNG with exceptional multi-constraint adherence.	Evidence A →

Capability · 95.5/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	64.5 / 65	Text rendering (25.0/25: Wooden sign prints 'REALTEST 2026' with 100% accurate spelling and engraved wood texture) · Count & color binding (20.0/20: Exactly 3 red apples on the left and 2 blue ceramic mugs on the right on a dark oak table) · Anatomy & structure (19.5/20: Brass robot's right hand clearly shows 4 mechanical fingers + 1 thumb gripping the top edge of the wooden sign).	Evidence A →
Deliverable specifications	20 / 20	Clean unwatermarked export (8/8); native export is a 2048×2048 (2K) RGB lossless PNG (8,158,105 bytes) exceeding the 1024px baseline (12/12).	Evidence A →
Processing speed	11 / 15	2026-10-21 same-run CDP network timing: 10.49s server wait from POST /api/jobs/v2/new/k3 to job completion notification POST /api/notifications/[id]/acknowledge (10.49s end-to-end, async_verified) (11.0/15).	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

applore.app

applore.app · Image generation

92/100

Seven-aspect task score

APPLORE demonstrates top-tier instruction adherence and rendering precision in the image generation benchmark. Powered by GPT Image 2.5, it achieved perfect scores across text rendering ("REALTEST 2026"), color/quantity binding (3 apples, 2 blue mugs), and 5-finger robot hand anatomy (65/65). New accounts receive 10 trial credits, and 5 credits were used to generate and export a native 1024x1024 unwatermarked PNG. With transparent credit deduction and an ~15s generation latency, it represents a highly reliable tool for icon and graphic design.

Tested export 1024 × 1024 · PNG

Historical test · 2026-10-28
Signed-in flow; paid tier not standardized · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (1024 × 1024 · PNG) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

100% accurate text spelling, zero color bleeding across specified objects, structurally sound 5-finger robotic hand, and native 1024x1024 unwatermarked PNG export.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

100% accurate text spelling, zero color bleeding across specified objects, structurally sound 5-finger robotic hand, and native 1024x1024 unwatermarked PNG export.

Original historical observation

Outstanding performance with GPT Image 2.5: exact text spelling, 3 apples and 2 mugs with zero color bleeding, sound 5-finger robot hand anatomy, and native 1024x1024 PNG export (2.55MB).

Suitable uses · historical assessment

Designers and developers needing accurate typography rendering, multi-entity spatial/color binding, and clean 1024px icon assets

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

1024 × 1024 · PNG · No watermark in test

Detailed facts & historical record

Recorded export specifications: 1024 × 1024 · PNG · No watermark in test
Historical specification note · not an official maximum: 1024×1024 (PNG)
Watermark observation: Completely watermark-free (1024x1024 native PNG export)
Batch & API: Configurable 1-4 icons per run, integrated with asset viewer, prompt reuse, and logo animation links
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: 1024 × 1024 · PNG

Evidence & scope →

Speed & reliability

About 15.5s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 15s · End-to-end reference: 15.5s
Date & fixture: 2026-10-28 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Supports Google OAuth and email sign-in (10 free trial credits upon registration)
Manual adjustment: Not verified
Interaction observations: Signups receive 10 trial credits (5 per icon generation, allowing 2 full runs), supporting 1024x1024 unwatermarked PNG export and multi-size previews without invasive ads or dark patterns.

Evidence & scope →

Background & reputation

APPLORE Team · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: APPLORE Team
Recorded origin: Not verified
Domain history record: Not verified
Background & reputation assessment: Independent utility platform with bilingual interface, clear Terms and Privacy policies, and dedicated support email.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Google Analytics · Stripe Checkout
Engineering & privacy assessment: Exports 1024x1024 unwatermarked PNG (2.55MB) via cloud inference with async polling and status queries; average inference time ~15.0s.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

APPLORE sitewide icon archive and AI credits; free trial, monthly subscriptions, and lifetime access · 2026-10-28

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free Tier2026-10-28 ↗

Free

Free 25 AI credits on signup, 5 daily downloads, browse first 300 icons per category

Pro Monthly2026-10-28 ↗

USD 9.99/month

—

Monthly 500 monthly AI credits (~100 icons), full 18,700+ icon archive with redesign history and SVG export

Max Monthly2026-10-28 ↗

USD 29.99/month

—

Monthly 2,000 monthly AI credits (~400 icons), 8 concurrent icons & 6 animations, priority queue

Founder Lifetime2026-10-28 ↗

USD 49 · one-time payment

Usage / one-time

Usage / one-time One-time lifetime access to full icon library and non-AI tools, with 800 included AI credits

4 plan / billing-period records, all listed; — means that period is not recorded.

25 free trial credits upon signup; non-AI tools and basic browsing are free; Pro includes 500 monthly credits and full archive; Founder tier provides lifetime access.

Cancellation, sources & verification record

Cancellation and refund terms unverified

APPLORE official homepage and icon explorer ↗ · 2026-10-28 Capture 1 ↗

APPLORE official pricing plans page ↗ · 2026-10-28 Capture 1 ↗

Allowance & disclosure on test date

10 trial credits upon signup; 5 credits per generation, allowing 2 full free runs

5 credits clearly indicated next to Create button with real-time remaining credit counter

Evaluation basis & full scores · Overall 92.3/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 86/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	33 / 35	Transparent pricing detailing trial tier, Pro ($9.99/mo), and Founder lifetime ($49) with clear per-feature credit deduction table.	Evidence A →
Interaction restraint	26 / 30	Signups receive 10 trial credits (5 per icon generation, allowing 2 full runs), supporting 1024x1024 unwatermarked PNG export and multi-size previews without invasive ads or dark patterns.	Evidence A →
Background & reputation	14 / 20	Independent utility platform with bilingual interface, clear Terms and Privacy policies, and dedicated support email.	Evidence A →
Engineering & privacy	13 / 15	Exports 1024x1024 unwatermarked PNG (2.55MB) via cloud inference with async polling and status queries; average inference time ~15.0s.	Evidence A →

Capability · 95/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	65 / 65	Text Rendering (25/25: Wooden sign with engraved 'REALTEST 2026' 100% spelled correctly, sharp and centered (25/25)) · Count & Color Binding (20/20: Dark oak table accurately displays 3 red apples and 2 blue ceramic mugs with zero color bleeding (20/20)) · 5-Finger Hand & Structure (20/20: Vintage brass robot right hand holds wooden sign with 5 distinct articulated fingers (20/20))	Evidence A →
Deliverable specifications	20 / 20	Forced watermark in free export (0/8); Export & Max deliverable spec: 1024×1024 (PNG)	Evidence A →
Processing speed	10 / 15	Network-observed server wait: 15s (API mode: async_poll; end-to-end 15.5s shown for reference only, never penalized)	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Pixlr

pixlr.com · Image generation

92/100

Seven-aspect task score

Upfront credit costs and preserved prompts after login accompany good text, counts and fast generation; an account and limited free credits are required.

Tested export 1024×1024 (Unwatermarked PNG, 1.38MB) · Historical description; output not attached

Historical test · 2026-10-16
Signed-in flow; paid tier not standardized · Fixed English prompt · Historical page record; output not attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Text and object counts were accurate for this fixed prompt; findings do not cover other prompts.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Text and object counts were accurate for this fixed prompt; findings do not cover other prompts.

Original historical observation

Major reverse feature bias compared to its weak Background Remover (63.7): Pixlr's AI Image Generator ranks #1 overall (92.1) by pairing FLUX.2 Klein 9B with 10 free 1-credit generations.

Suitable uses · historical assessment

Users wanting the highest overall free T2I experience: fast (4.6s), accurate typography (FLUX.2 Klein 9B), and 10 free unwatermarked 1024×1024 exports upon login

Result & output evidence

No native output obtained; the preview does not establish final export quality.

Open result screenshot ↗

Full test record →

Evidence & scope →

Delivery capabilities

1024×1024 (Unwatermarked PNG, 1.38MB) · Historical description; output not attached · Check watermark record

Detailed facts & historical record

Recorded export specifications

1024×1024 (Unwatermarked PNG, 1.38MB) · Historical description; output not attached · Check watermark record

No native output is attached; check the evidence for specification and watermark claims.

Historical specification note · not an official maximum

1024×1024 (Unwatermarked PNG · FLUX.2 Klein 9B)

Watermark observation

Zero watermark (clean 1024×1024 PNG export)

Batch & API

Supports 1–4 images per prompt and multiple model choices (FLUX.2, Nano Banana, Seedream, Ideogram)

Recorded free-tier dimensions

Not verified

Attached file dimensions

Not verified

Evidence & scope →

Speed & reliability

About 5.4s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 4.6s · End-to-end reference: 5.4s
Date & fixture: 2026-10-16 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Free Google / Apple / Email login (grants 10 free initial credits)
Manual adjustment: Not verified
Interaction observations: Google OAuth login returns directly to the generator page with the user's typed prompt 100% preserved—zero prompt loss, zero forced paywalls, and direct one-click PNG download.

Evidence & scope →

Background & reputation

Pixlr Pte. Ltd. (Inmagine Group) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Pixlr Pte. Ltd. (Inmagine Group)
Recorded origin: Singapore / Malaysia (founded in Sweden 2008)
Domain history record: 2008–present
Background & reputation assessment: Operated by Pixlr Pte. Ltd. (Inmagine Group, founded 2008), a veteran browser creative platform.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Defaults to the fast FLUX.2 Klein 9B model, generating a 1024×1024 unwatermarked PNG in 4.6s Network-observed server wait.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

All web monthly/annual tiers and five credit top-up packs · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Plus2026-10-20 ↗

USD 2.49/month

USD 23.88/year

Monthly 80 AI credits/month; not a dedicated removal count Rules [1]

Annual 960 general AI credits/year, valid for one year Rules [2]

Premium2026-10-20 ↗

USD 9.99/month

USD 95.88/year

Monthly 1,000 AI credits/month; not a dedicated removal count Rules [1]

Annual 12,000 general AI credits/year, valid for one year Rules [2]

Ultra regular2026-10-20 ↗

USD 24.99/month

USD 239.9/year

Monthly 5,000 AI credits/month; not a dedicated removal count Rules [1]

Annual 60,000 general AI credits/year, valid for one year Rules [2]

Ultra MAX2026-10-20 ↗

USD 49.99/month

USD 479.9/year

Monthly 10,000 AI credits/month; not a dedicated removal count Rules [1]

Annual 120,000 general AI credits/year, valid for one year Rules [2]

200 AI credits2026-10-20 ↗

USD 3.99 · one-time payment

Usage / one-time

Usage / one-time 200 general AI credits; pack expiry not separately verified

500 AI credits2026-10-20 ↗

USD 6.99 · one-time payment

Usage / one-time

Usage / one-time 500 general AI credits; pack expiry not separately verified

1,000 AI credits2026-10-20 ↗

USD 9.99 · one-time payment

Usage / one-time

Usage / one-time 1,000 general AI credits; pack expiry not separately verified

2,000 AI credits2026-10-20 ↗

USD 16.99 · one-time payment

Usage / one-time

Usage / one-time 2,000 general AI credits; pack expiry not separately verified

5,000 AI credits2026-10-20 ↗

USD 30.99 · one-time payment

Usage / one-time

Usage / one-time 5,000 general AI credits; pack expiry not separately verified

13 plan / billing-period records, all listed; — means that period is not recorded.

[1] Credits valid for 1 month

[2] One annual payment; tax is calculated after billing address entry. Amount comes from checkout, not multiplication of rounded monthly equivalents.

All four annual charges were confirmed at checkout. Ultra $239.90/year and Ultra MAX $479.90/year differ from rounded monthly equivalents multiplied by 12. The feature FAQ advertises free removal and batch ZIP export without quantifying free counts/limits. Removal credit conversion and top-up expiry are not separately stated in reviewed pages. Tax depends on billing address.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Page states cancellation is available anytime; refund details unverified.

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗

Official annual checkout summaries for all four plans ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗ · Capture 4 ↗

Official free removal and batch explanation ↗ · 2026-10-20 Capture 1 ↗

Allowance & disclosure on test date

10 free credits upon signup; default FLUX.2 Klein 9B model costs only 1 credit per image (10 free exports)

Upfront credit cost ("Requires 1 credits per image") shown before clicking Generate

Evaluation basis & full scores · Overall 92.2/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 82/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	25 / 35	Landing page (01_landing.png) explicitly states "Requires 1 credits per image" and displays the button "Generate (1)"; clicking it opens a free login modal (01_pass1_login_modal.png), and Google sign-in immediately grants 10 free credits (enough for 10 free 1024×1024 unwatermarked generations).	Evidence A →
Interaction restraint	25 / 30	Google OAuth login returns directly to the generator page with the user's typed prompt 100% preserved—zero prompt loss, zero forced paywalls, and direct one-click PNG download.	Evidence A →
Background & reputation	18 / 20	Operated by Pixlr Pte. Ltd. (Inmagine Group, founded 2008), a veteran browser creative platform.	Evidence C →
Engineering & privacy	14 / 15	Defaults to the fast FLUX.2 Klein 9B model, generating a 1024×1024 unwatermarked PNG in 4.6s Network-observed server wait.	Evidence A →

Capability · 96.5/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	63.5 / 65	Text Rendering (25/25: Single-line wood-carved "REALTEST 2026" on the oak sign with 100% spelling accuracy and realistic engraving depth (25/25)) · Count & Color Binding (20/20: Strictly 3 glossy red apples on the left + 2 speckled blue ceramic mugs on the right, zero color bleeding (20/20)) · 5-Finger Hand & Structure (18.5/20: Detailed vintage brass robotic hands gripping both sides of the wooden sign with distinct mechanical finger joints (18.5/20))	Evidence B →
Deliverable specifications	20 / 20	Clean unwatermarked export (8/8); Export & Max deliverable spec: 1024×1024 (Unwatermarked PNG · FLUX.2 Klein 9B)	Evidence B →
Processing speed	13 / 15	Network-observed server wait: 4.6s (API mode: sync; end-to-end 5.4s shown for reference only, never penalized)	Evidence B →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

befunky.com

befunky.com · Image generation

91/100

Seven-aspect task score

BeFunky offers 20 free unwatermarked AI generations, delivering a native 2048×2048 export in just 3.17s server wait with excellent 5-finger anatomy and rich ceramic textures; sign lettering has minor suffix noise (REALTEST ST), while background removal and upscaling require Plus.

Tested export JPEG · Output attached

Historical test · 2026-11-01
Signed-in flow; paid tier not standardized · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (JPEG · Output attached) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Vintage brass robot exhibits 5 distinct mechanical fingers and dual blue mugs with exquisite texture (20/20); sign text includes extra ST suffix (18/25) and generates 2 red apples (16/20) for 54.0/65.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Original historical observation

BeFunky AI Image Generator combines fast 3.17s server speed with high-res 2048×2048 output, realistic brass reflections, and 5-finger anatomy, with minor deviations on apple count (2 vs 3) and sign lettering suffix.

Suitable uses · historical assessment

Creators wanting fast 2048px image generation integrated into a full-featured online photo and graphic design suite

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

JPEG · Output attached · Check watermark record

Detailed facts & historical record

Recorded export specifications: JPEG · Output attached · Check watermark record
Historical specification note · not an official maximum: 2048×2048 (JPEG/PNG)
Watermark observation: Completely free of visible watermarks
Batch & API: Supports generation history tab, 1:1 aspect ratio, and seamless handoff into BeFunky Photo Editor
Recorded free-tier dimensions: 2048 × 2048
Attached file dimensions: Not verified

Evidence & scope →

Speed & reliability

About 4.93s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 3.17s · End-to-end reference: 4.93s
Date & fixture: 2026-11-01 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Supports free Google / Facebook / Apple / Email sign-in
Manual adjustment: Not verified
Interaction observations: Studio workspace provides an integrated Image Generator modal supporting 1:1 Square aspect ratio and 1-10 Variation controls, with direct Save to Computer options (JPG/PNG/PDF) and zero disruptive paywall traps.

Evidence & scope →

Background & reputation

BeFunky Inc. · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: BeFunky Inc.
Recorded origin: Not verified
Domain history record: Not verified
Background & reputation assessment: Inherited from BeFunky product baseline: Founded in 2007 with nearly 20 years of continuous operation, verifiable leadership, and strong commercial reputation.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Google Analytics · Facebook Pixel
Engineering & privacy assessment: Server inference wait measured at 3.17s (POST mapi.befunky.com/t2i/generate -> GET gpu-nrck-rlg1.befunky.ai/uploads/...), exporting an unwatermarked 2048×2048 (1.12MB) JPEG with BeFunky Photo Editor Exif metadata.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

BeFunky Free, Plus, and Pro monthly and annual subscriptions · 2026-11-01

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

BeFunky Free2026-11-01 ↗

Free

Free Up to 20 AI image generations with free account, basic editing

Basic photo editing tools, limited templates, export with no subscription；Advanced AI tools (background removal, upscaling) and premium assets locked；Free forever, no renewal；Perpetual access

BeFunky Plus Monthly2026-11-01 ↗

USD 14.99/month

—

Monthly 1,000 AI credits monthly for AI Image Generator, top-up anytime

Full access to premium editing and design tools, AI editing controls, Artsy effects, bulk editing

Rules [1] [2] [3]

BeFunky Plus Annual2026-11-01 ↗

—

USD 83.88/year

Annual 1,000 AI credits monthly ($6.99/mo equivalent billed $83.88 annually)

Full access to premium editing and design tools, AI editing controls, Artsy effects, bulk editing；$83.88 billed upfront annually

Rules [4] [5]

BeFunky Pro Monthly2026-11-01 ↗

USD 24.99/month

—

Monthly 5,000 AI credits monthly, priority processing queue

All Plus features plus Batch AI matting, upscaling/restoration at scale, advanced models, priority support

Rules [1] [2] [3]

BeFunky Pro Annual2026-11-01 ↗

—

USD 149.88/year

Annual 5,000 AI credits monthly ($12.49/mo equivalent billed $149.88 annually)

All Plus features plus Batch AI matting, upscaling/restoration at scale, advanced models, priority support；$149.88 billed upfront annually

Rules [4] [5]

5 plan / billing-period records, all listed; — means that period is not recorded.

[1] Billed monthly

[2] Monthly automatic renewal, cancel anytime

[3] 1 month validity

[4] Annual automatic renewal, cancel anytime

[5] 1 year validity

Free tier covers basic edits and 20 AI generations; Plus is $14.99/mo or $83.88/yr ($6.99/mo eq, 1,000 credits/mo); Pro is $24.99/mo or $149.88/yr ($12.49/mo eq, 5,000 credits/mo and batch AI).

Cancellation, sources & verification record

Cancellation and refund terms unverified

BeFunky Annual Plans Pricing ↗ · 2026-11-01 Capture 1 ↗

BeFunky Monthly Plans Pricing ↗ · 2026-11-01 Capture 1 ↗

Allowance & disclosure on test date

20 free AI image generations on signup with visible remaining counter

Counter displays '20 FREE IMAGES REMAINING' prominently before and after generation

Evaluation basis & full scores · Overall 90.8/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 95/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	34 / 35	Inherited from BeFunky product baseline: Clear pricing disclosures across Free ($0), Plus ($14.99/mo or $83.88/yr), and Pro ($24.99/mo or $149.88/yr, 5,000 credits).	Evidence A →
Interaction restraint	28 / 30	Studio workspace provides an integrated Image Generator modal supporting 1:1 Square aspect ratio and 1-10 Variation controls, with direct Save to Computer options (JPG/PNG/PDF) and zero disruptive paywall traps.	Evidence A →
Background & reputation	18 / 20	Inherited from BeFunky product baseline: Founded in 2007 with nearly 20 years of continuous operation, verifiable leadership, and strong commercial reputation.	Evidence A →
Engineering & privacy	15 / 15	Server inference wait measured at 3.17s (POST mapi.befunky.com/t2i/generate -> GET gpu-nrck-rlg1.befunky.ai/uploads/...), exporting an unwatermarked 2048×2048 (1.12MB) JPEG with BeFunky Photo Editor Exif metadata.	Evidence A →

Capability · 89/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	54 / 65	Text rendering (18.0/25: Sign renders 'REALTEST ST 2026' with minor trailing ST redundancy) · Count & color binding (16.0/20: Exactly 2 blue ceramic mugs and 2 red apples on oak table without bleeding) · Structure & composition (20.0/20: 5 distinct mechanical joints on the brass robot right hand holding the wooden sign).	Evidence A →
Deliverable specifications	20 / 20	Clean unwatermarked export (8/8); native deliverable reaches 2048×2048 full-resolution JPEG (12/12).	Evidence A →
Processing speed	15 / 15	Verified server processing time 3.17s (fast < 4.0s, async_verified, 4.93s end-to-end) (15.0/15).	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Nero AI

ai.nero.com · Image generation

91/100

Seven-aspect task score

Free signup credits run Basic Nano Banana (1:1, 1K, 4 credits/image) to export a 1024×1024 unwatermarked PNG in 7.0s server wait with exact sign text and 5-digit hands; requires sign-in and rendered one extra apple.

Tested export 1024 × 1024 · PNG

Historical test · 2026-10-21
Signed-in flow; paid tier not standardized · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (1024 × 1024 · PNG) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

'REALTEST 2026' sign lettering (25/25) and 5-digit robot hand structure (19/20) are accurate; right side shows 2 blue mugs accurately while left side rendered 4 red apples instead of 3 (60.0/65).

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

'REALTEST 2026' sign lettering (25/25) and 5-digit robot hand structure (19/20) are accurate; right side shows 2 blue mugs accurately while left side rendered 4 red apples instead of 3 (60.0/65).

Original historical observation

Google test account generated canonical 1:1 prompt using Basic Nano Banana, 4 credits. Downloaded original result. Signup showed 10 base plus 20 newsletter credits; accidental newsletter opt-in is being reversed and bonus excluded from test-budget accounting.

Suitable uses · historical assessment

Users wanting fast 1K text-accurate illustrations and anatomical hand structure via Basic Nano Banana

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

1024 × 1024 · PNG · No watermark in test

Detailed facts & historical record

Recorded export specifications: 1024 × 1024 · PNG · No watermark in test
Historical specification note · not an official maximum: [1024, 1024]
Watermark observation: No watermark (native 1024×1024 lossless RGB PNG)
Batch & API: Conversational agent interface supporting Basic Nano Banana (1K) and Pro Nano Banana (2K/4K) with multi-turn editing
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: 1024 × 1024 · PNG

Evidence & scope →

Speed & reliability

About 7.51s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 7.01s · End-to-end reference: 7.51s
Date & fixture: 2026-10-21 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Authorized test profile; see dated run record
Manual adjustment: Not verified
Interaction observations: Historical rubric scores withdrawn pending review of the new run. Google test account generated canonical 1:1 prompt using Basic Nano Banana, 4 credits. Downloaded original result. Signup showed 10 base plus 20 newsletter credits; accidental newsletter opt-in is being reversed and bonus excluded from test-budget accounting.

Evidence & scope →

Background & reputation

Established German Multimedia Software Vendor (Nero AG, est. 1995) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Established German Multimedia Software Vendor (Nero AG, est. 1995)
Recorded origin: Karlsruhe, Germany
Domain history record: 1995–present
Background & reputation assessment: Historical rubric scores withdrawn pending review of the new run. Google test account generated canonical 1:1 prompt using Basic Nano Banana, 4 credits. Downloaded original result. Signup showed 10 base plus 20 newsletter credits; accidental newsletter opt-in is being reversed and bonus excluded from test-budget accounting.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Historical rubric scores withdrawn pending review of the new run. Google test account generated canonical 1:1 prompt using Basic Nano Banana, 4 credits. Downloaded original result. Signup showed 10 base plus 20 newsletter credits; accidental newsletter opt-in is being reversed and bonus excluded from test-budget accounting.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Nero AI web plans (Free / Standard / Yearly promo / Business); USD · 2026-10-21

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free2026-10-21 ↗

Free

Free 5 free guest runs without login; 10 one-time credits per registered account (max 1000×1000px upload, 5 images/batch; 2× costs 1 credit, 4× costs 2 credits).

Standard Monthly (300 credits)2026-10-21 ↗

USD 21.95/month

—

Monthly 300 credits/month (promotional USD 21.95/month, regular USD 24.99/month); up to 4000×4000px upload, 30 images/batch, priority processing, and 8× upscaling.

Standard Yearly (3,000 credits/yr)2026-10-21 ↗

—

Annual price not verified

Annual 3,000 credits/year; promotional banner displays USD 7.66/month equivalent.

Public page displays only the 7.66 USD/month equivalent without showing the annual invoice total before login.

Business2026-10-21 ↗

Contact sales

Quoted

Quoted Custom credits and pricing (up to 200 images/batch, highest priority, API processing, and dedicated server).

4 plan / billing-period records, all listed; — means that period is not recorded.

The official pricing page and credit table state that guests get 5 free runs and registered Free accounts receive 10 one-time Nero Credits (Standard/Photo/Anime 2× costs 1 credit/image, 4× costs 2 credits/image, paid-only 8× costs 4 credits/image). Standard monthly is USD 21.95/month promotional (regular USD 24.99/month, 300 credits/month). The yearly promo banner displays only "7.66 USD per month" (3,000 credits/year) without printing the annual invoice total before login, so the annual total is marked unverified.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Cancel anytime by turning off auto-renewal in account settings (remaining credits expire at the end of the billing period); FAQ states refunds are not allowed once Nero Credits have been used.

Nero AI official pricing, credit consumption table, and FAQ ↗ · 2026-10-21 Capture 1 ↗

Allowance & disclosure on test date

Submit button explicitly shows the 4-credit cost and top header displays remaining free credits before submission

Evaluation basis & full scores · Overall 90.5/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 87/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	28 / 35	Historical rubric scores withdrawn pending review of the new run. Google test account generated canonical 1:1 prompt using Basic Nano Banana, 4 credits. Downloaded original result. Signup showed 10 base plus 20 newsletter credits; accidental newsletter opt-in is being reversed and bonus excluded from test-budget accounting.	Evidence A →
Interaction restraint	26 / 30	Historical rubric scores withdrawn pending review of the new run. Google test account generated canonical 1:1 prompt using Basic Nano Banana, 4 credits. Downloaded original result. Signup showed 10 base plus 20 newsletter credits; accidental newsletter opt-in is being reversed and bonus excluded from test-budget accounting.	Evidence A →
Background & reputation	19 / 20	Historical rubric scores withdrawn pending review of the new run. Google test account generated canonical 1:1 prompt using Basic Nano Banana, 4 credits. Downloaded original result. Signup showed 10 base plus 20 newsletter credits; accidental newsletter opt-in is being reversed and bonus excluded from test-budget accounting.	Evidence A →
Engineering & privacy	14 / 15	Historical rubric scores withdrawn pending review of the new run. Google test account generated canonical 1:1 prompt using Basic Nano Banana, 4 credits. Downloaded original result. Signup showed 10 base plus 20 newsletter credits; accidental newsletter opt-in is being reversed and bonus excluded from test-budget accounting.	Evidence A →

Capability · 92/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	60 / 65	Google test account generated canonical 1:1 prompt using Basic Nano Banana, 4 credits. Downloaded original result. Signup showed 10 base plus 20 newsletter credits; accidental newsletter opt-in is being reversed and bonus excluded from test-budget accounting.	Evidence A →
Deliverable specifications	20 / 20	Native file obtained; see format/dimension inspection. Google test account generated canonical 1:1 prompt using Basic Nano Banana, 4 credits. Downloaded original result. Signup showed 10 base plus 20 newsletter credits; accidental newsletter opt-in is being reversed and bonus excluded from test-budget accounting.	Evidence A →
Processing speed	12 / 15	Network-observed server wait was not reliably isolated; speed and aggregate scores remain unknown.	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Adobe

adobe.com · Image generation

90/100

Seven-aspect task score

Free Adobe daily allowance runs Firefly Image 5 (1:1, 1K) to export a 1024×1024 unwatermarked PNG in 6.21s server wait with exact sign text and object counts, though the robot hand uses a 3-ring clamp; login and daily limits apply.

Tested export 1024 × 1024 · PNG

Historical test · 2026-10-21
Signed-in flow; paid tier not standardized · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (1024 × 1024 · PNG) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

'REALTEST 2026' lettering and 3 red apples + 2 blue ceramic mugs are 100% accurate; the retro robot hand is stylized as a 3-ring clamp rather than 5 distinct fingers (59.0/65).

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

'REALTEST 2026' lettering and 3 red apples + 2 blue ceramic mugs are 100% accurate; the retro robot hand is stylized as a 3-ring clamp rather than 5 distinct fingers (59.0/65).

Original historical observation

Strong typography and object counting (59.0/65): 'REALTEST 2026' text and 3 red apples + 2 blue ceramic mugs are 100% exact, while the retro robot hand is stylized as a 3-segment brass clamp instead of 5 distinct fingers.

Suitable uses · historical assessment

Commercial creators needing copyright-safe C2PA imagery, exact sign typography, and seamless handoff to Photoshop Web

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

1024 × 1024 · PNG · Check watermark record

Detailed facts & historical record

Recorded export specifications: 1024 × 1024 · PNG · Check watermark record
Historical specification note · not an official maximum: 1024×1024 (1K PNG)
Watermark observation: No visible watermark (native 1024×1024 RGB PNG with C2PA metadata)
Batch & API: Supports composition reference upload, mosaic/carousel history view, and one-click opening in Photoshop Web
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: 1024 × 1024 · PNG

Evidence & scope →

Speed & reliability

About 6.82s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 6.21s · End-to-end reference: 6.82s
Date & fixture: 2026-10-21 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Supports free Google / Apple / Facebook / Email Adobe account sign-in
Manual adjustment: Not verified
Interaction observations: Left inspector clearly exposes Model (Firefly Image 5), Aspect Ratio (1:1), and Resolution (1K); hovering the generated image provides one-click Download or Open in Photoshop Web. A dismissible 30%-off banner appears below the canvas without blocking export.

Evidence & scope →

Background & reputation

Adobe Inc. (NASDAQ: ADBE) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Adobe Inc. (NASDAQ: ADBE)
Recorded origin: San Jose, CA, USA
Domain history record: Founded 1982 (domain registered 1986)
Background & reputation assessment: Flagship Firefly Image 5 model by Adobe Inc. (NASDAQ: ADBE), trained on licensed Adobe Stock content with built-in C2PA Content Credentials.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Dispatches async generation via POST image-v5.ff.adobe.io/v1/images/generate-async and polls bks-epo8552.adobe.io/v2/jobs/result (network-observed server wait 6.21s), delivering a native 1024×1024 lossless RGB PNG.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

US individual Firefly web free feature and four monthly subscriptions; excludes student/business, full Creative Cloud and other regions; includes four generative credit add-on subscriptions · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Firefly web background remover2026-10-20 ↗

Free

Free Official feature page describes free removal; single JPG/PNG/WebP input up to 50MB

Count and free export specification are not stated; subsequent generative editing may require separate credits

Firefly Standard2026-10-20 ↗

USD 9.99/month

—

Monthly 2,000 generative credits/month; not removal counts Rules [1]

Firefly Pro2026-10-20 ↗

USD 19.99/month

—

Monthly 4,000 generative credits/month; not removal counts Rules [1]

Firefly Pro Plus2026-10-20 ↗

USD 34.97/month

—

Monthly 10,000 generative credits/month; not removal counts

First-year promotion advertised through October 21; regular price USD 49.99/month. Renewal and eligibility follow the offer terms.

Rules [1]

Firefly Premium2026-10-20 ↗

USD 139.91/month

—

Monthly 50,000 generative credits/month; not removal counts

First-year promotion advertised through October 21; regular price USD 199.99/month. Renewal and eligibility follow the offer terms.

Rules [1]

2,000 generative credit add-on2026-10-20 ↗

USD 9.99/month

—

Monthly 2,000 generative credits/month; not removal counts

7,000 generative credit add-on2026-10-20 ↗

USD 29.99/month

—

Monthly 7,000 generative credits/month; not removal counts

10,000 generative credit add-on2026-10-20 ↗

USD 49.99/month

—

Monthly 10,000 generative credits/month; not removal counts

50,000 generative credit add-on2026-10-20 ↗

USD 199.99/month

—

Monthly 50,000 generative credits/month; not removal counts

9 plan / billing-period records, all listed; — means that period is not recorded.

[1] US individual web price; the removal feature itself is advertised as free, so a generative subscription is not established as a removal requirement

Four US individual Firefly monthly plans and four credit add-on subscriptions are listed. Pro Plus/Premium first-year promotions end October 21, with regular prices separately shown. Official pages do not quantify daily free removal counts or regional tax amounts. Generative credits concern other AI features and do not establish a payment requirement for free removal.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Self-serve cancellation in Adobe Account (monthly plans cancel anytime; annual billed-monthly plans subject to Adobe terms)

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗

US individual Firefly subscriptions ↗ · 2026-10-20 Capture 1 ↗

Official generative credit add-on subscriptions ↗ · 2026-10-20 Capture 1 ↗

Allowance & disclosure on test date

Free Adobe account includes daily free generations ('Daily free generations' badge in prompt bar) supporting Firefly Image 5 (1K)

Prompt bar explicitly displays remaining daily free generations before and after each run

Evaluation basis & full scores · Overall 90.1/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 88/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	28 / 35	Landing page provides a prompt entry box and 'Go to Firefly' link; signing in with a free Adobe account on firefly.adobe.com/generate/images displays 'Daily free generations (2 left today)' and exports a native 1024×1024 unwatermarked PNG (1.41MB) using Firefly Image 5 (1:1, 1K).	Evidence A →
Interaction restraint	26 / 30	Left inspector clearly exposes Model (Firefly Image 5), Aspect Ratio (1:1), and Resolution (1K); hovering the generated image provides one-click Download or Open in Photoshop Web. A dismissible 30%-off banner appears below the canvas without blocking export.	Evidence A →
Background & reputation	20 / 20	Flagship Firefly Image 5 model by Adobe Inc. (NASDAQ: ADBE), trained on licensed Adobe Stock content with built-in C2PA Content Credentials.	Evidence A →
Engineering & privacy	14 / 15	Dispatches async generation via POST image-v5.ff.adobe.io/v1/images/generate-async and polls bks-epo8552.adobe.io/v2/jobs/result (network-observed server wait 6.21s), delivering a native 1024×1024 lossless RGB PNG.	Evidence A →

Capability · 91/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	59 / 65	Text rendering (25.0/25: Wooden sign prints 'REALTEST 2026' across two lines with 100% accurate spelling) · Count & color binding (20.0/20: Exactly 3 red apples on the left and 2 blue ceramic mugs on the right on a dark oak table) · Anatomy & structure (14.0/20: Vintage brass robot's right arm holds the sign stick with a 3-ring + thumb clamp claw rather than 5 distinct humanoid fingers).	Evidence A →
Deliverable specifications	20 / 20	Clean unwatermarked export (8/8); native export is a 1024×1024 (1K) RGB lossless PNG (1,412,597 bytes) with C2PA Content Credentials (12/12).	Evidence A →
Processing speed	12 / 15	2026-10-21 same-run CDP network timing: 6.21s server wait from POST image-v5.ff.adobe.io/v1/images/generate-async to terminal GET bks-epo8552.adobe.io/v2/jobs/result poll (6.82s end-to-end, async_verified) (12.0/15).	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Vheer

vheer.com · Image generation

89/100

Seven-aspect task score

Free login-free 1248px clean downloads are not blocked by ads; video ads remain on the page and this result did not fully match requested object counts.

Tested export PNG · Output attached

Historical test · 2026-10-17
Guest flow · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (PNG · Output attached) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Delivered 1248px without login with correct text in this test; other prompts remain untested.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Delivered 1248px without login with correct text in this test; other prompts remain untested.

Original historical observation

Vheer AI Image Generator spells "REALTEST CAFE" with 100% letter accuracy on the hanging wooden sign, renders rich warm bokeh lighting, and places the brass microscope and colored espresso cups cleanly at 1248×1248 resolution.

Suitable uses · historical assessment

Creators seeking a 100% free, no-login Flux text-to-image generator with accurate English typography rendering and 1248×1248 unwatermarked exports

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

PNG · Output attached · No watermark in test

Detailed facts & historical record

Recorded export specifications: PNG · Output attached · No watermark in test
Historical specification note · not an official maximum: 1248 × 1248 (unwatermarked lossless PNG)
Watermark observation: Watermark-free
Batch & API: Supports multiple aspect ratios (1:1 1248×1248, 16:9, 9:16, etc.) and rapid regeneration without login
Recorded free-tier dimensions: 1248 × 1248
Attached file dimensions: Not verified

Evidence & scope →

Speed & reliability

About 11.7s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 11.5s · End-to-end reference: 11.7s
Date & fixture: 2026-10-17 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

No login in this run

Detailed facts & historical record

Account requirements: No login in this test
Sign-in methods: 100% usable without login
Manual adjustment: Not verified
Interaction observations: No registration modal, no email gate, and no watermark trap; supports aspect-ratio selection and iterative regeneration on the same page.

Evidence & scope →

Background & reputation

Vheer AI Creative Suite · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Vheer AI Creative Suite
Recorded origin: AI visual product team behind ImgUpscaler & Vheer
Domain history record: 2024–Present
Background & reputation assessment: Flagship creative AI hub from the ImgUpscaler / BgEraser team, integrating Flux-based text-to-image and vision tools.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: HTTPS encrypted generation without requiring personal account binding or persistent prompt history tracking.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Plans with sourced commercial rights: Premium / Business / Ultra

Website free, monthly, annual and general credit packs · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free2026-10-20 ↗

Free

Free Removal feature page states free use without signup/subscription, up to 20 images per batch; generation separately includes 50 general credits/month

Premium2026-10-20 ↗

USD 9.99/month

USD 99.9/year

Monthly 1000 general credits/month, commercial use, no ads/watermark

Annual 1000 general credits/month; displayed monthly equivalent $8.33 Rules [1]

Business2026-10-20 ↗

USD 19.99/month

USD 199.9/year

Monthly 2000 general credits/month, commercial use, no ads/watermark

Annual 2000 general credits/month; displayed monthly equivalent $16.66 Rules [1]

Ultra2026-10-20 ↗

USD 49.99/month

USD 499.9/year

Monthly 5000 general credits/month, commercial use, no ads/watermark

Annual 5000 general credits/month; displayed monthly equivalent $41.66 Rules [1]

400 credit pack2026-10-20 ↗

USD 3.99 · one-time payment

Usage / one-time

Usage / one-time 400 base general credits; higher-plan bonuses not quantified

950 credit pack2026-10-20 ↗

USD 9.49 · one-time payment

Usage / one-time

Usage / one-time 950 base general credits; higher-plan bonuses not quantified

1850 credit pack2026-10-20 ↗

USD 18.49 · one-time payment

Usage / one-time

Usage / one-time 1850 base general credits; higher-plan bonuses not quantified

2700 credit pack2026-10-20 ↗

USD 26.99 · one-time payment

Usage / one-time

Usage / one-time 2700 base general credits; higher-plan bonuses not quantified

3500 credit pack2026-10-20 ↗

USD 34.99 · one-time payment

Usage / one-time

Usage / one-time 3500 base general credits; higher-plan bonuses not quantified

7000 credit pack2026-10-20 ↗

USD 69.99 · one-time payment

Usage / one-time

Usage / one-time 7000 base general credits; higher-plan bonuses not quantified

13 plan / billing-period records, all listed; — means that period is not recorded.

[1] Official annual billing tooltip verified; checkout requires login, and tax/regional differences are unconfirmed

Annual prices come from official billing tooltips; upgrade checkout requires login. The removal feature explicitly states free use without registration and up to 20 images per batch, so generation credits are not converted into removal counts. Higher-tier top-up bonus quantities and exact credit expiry dates are not specified in the public rules.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

The FAQ allows cancellation anytime and says reset credits remain usable; the exact expiry is unspecified.

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗

Official annual billing details ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗

Official removal feature and batch rules ↗ · 2026-10-20 Capture 1 ↗

Allowance & disclosure on test date

Unlimited free generations without login or credit deductions

All aspect ratio and model options are immediately visible and unlocked on the workspace

Evaluation basis & full scores · Overall 88.5/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 90/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	32.5 / 35	Prominently advertises "Free AI Image Generator — No Sign-Up, Unlimited Generations" on the landing screen (01_landing.png). Tested without login: generates and exports 1248×1248 unwatermarked PNGs completely free with zero credit walls (02_workspace.png, 03_download.png).	Evidence A →
Interaction restraint	26.5 / 30	No registration modal, no email gate, and no watermark trap; supports aspect-ratio selection and iterative regeneration on the same page.	Evidence A →
Background & reputation	17.5 / 20	Flagship creative AI hub from the ImgUpscaler / BgEraser team, integrating Flux-based text-to-image and vision tools.	Evidence A →
Engineering & privacy	13.5 / 15	HTTPS encrypted generation without requiring personal account binding or persistent prompt history tracking.	Evidence A →

Capability · 87.8/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	58 / 65	Text Rendering (25/25: Double-line carved black text "REALTEST 2026" on the wooden sign is 100% accurately spelled and crisp) · Count & Color Binding (16/20: Renders 2 red apples on the left + 2 light-blue ceramic mugs on the right with zero color bleeding) · 5-Finger Hand & Structure (17/20: Retro brass robot holds the wooden sign with 4 fingers + thumb clearly articulated on the right hand)	Evidence A →
Deliverable specifications	18.8 / 20	Exports a high-resolution 1248×1248 unwatermarked PNG for free without login (16.5/20)	Evidence A →
Processing speed	11 / 15	Completes 1248×1248 Flux text-to-image synthesis in ~11.5s without login queues (12.0/15)	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

MyEdit

myedit.online · Image generation

88/100

Seven-aspect task score

Free accounts can claim 3 daily credits and run Nano Banana 2 Lite (1:1, 1 image, 2 credits) to export an unwatermarked 1024×1024 JPEG in 7.7s with exact sign text and 5-digit hands; requires login and rendered one extra apple.

Tested export JPEG · Output attached

Historical test · 2026-10-21
Signed-in flow; paid tier not standardized · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (JPEG · Output attached) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

'REALTEST 2026' sign lettering (25/25) and 5-digit brass robot hand structure (20/20) are exact; 2 blue mugs behind are accurate while 4 red apples appear in front instead of 3 (61.0/65).

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

'REALTEST 2026' sign lettering (25/25) and 5-digit brass robot hand structure (20/20) are exact; 2 blue mugs behind are accurate while 4 red apples appear in front instead of 3 (61.0/65).

Original historical observation

Strong multi-constraint adherence (61.0/65): 'REALTEST 2026' wooden sign text (25/25) and 5-digit brass robot hand anatomy (20/20) are spot-on, with 2 blue ceramic mugs rendered accurately and 1 extra red apple (4 vs 3) in the foreground (16/20).

Suitable uses · historical assessment

Users who want watermark-free 1K text-to-image generation funded by daily free bonus credits

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

JPEG · Output attached · No watermark in test

Detailed facts & historical record

Recorded export specifications: JPEG · Output attached · No watermark in test
Historical specification note · not an official maximum: 1024×1024 (1K JPEG)
Watermark observation: No watermark (native 1024×1024 RGB JPEG export)
Batch & API: Supports Nano Banana 2 Lite, GPT Image 2.5, FLUX, and custom style references with 1–4 images per run
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: Not verified

Evidence & scope →

Speed & reliability

About 7.71s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 7.72s · End-to-end reference: 7.71s
Date & fixture: 2026-10-21 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: CyberLink email/password, Google, Facebook, or Apple login
Manual adjustment: Not verified
Interaction observations: Requires account sign-in before submitting a prompt, but model, aspect ratio, and image-count controls are clear, and clicking the download icon on the generated image directly saves a 1024×1024 unwatermarked JPEG without secondary upsell modals or watermarks.

Evidence & scope →

Background & reputation

CyberLink Corp. (TWSE: 5203) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: CyberLink Corp. (TWSE: 5203)
Recorded origin: New Taipei City, Taiwan
Domain history record: 2022–present (Parent founded 1996)
Background & reputation assessment: Developed and operated by CyberLink Corp. (New Taipei City, Taiwan, TWSE: 5203, founded 1996, maker of PowerDirector and PhotoDirector).
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Submits generation via POST https://myedit.online/tti/effect and retrieves the generated result via GET https://myedit.online/tti/effect/{id}/1 (server_s = 7.72s, end_to_end_s = 7.71s), exporting a native 1024×1024 unwatermarked RGB JPEG (191,938 bytes).
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Web image/video free, monthly, annual and both credit-pack classes; excludes audio-only subscriptions · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free2026-10-20 ↗

Free

Free Free watermark-free transparent PNG/blurred-background downloads; one new-background download/day

A free account is required to download; premium backgrounds require subscription. Daily transparent-cutout cap is not quantified

Creator Pro2026-10-20 ↗

USD 18/month

USD 84/year

Annual 500 credits/month

Monthly 500 general AI credits/month

Studio Pro2026-10-20 ↗

USD 22/month

USD 108/year

Annual 800 credits/month

Monthly 800 general AI credits/month

Free User Pack · 502026-10-20 ↗

USD 12 · one-time payment

Usage / one-time

Usage / one-time 50 general AI credits; total calculated as count × published unit price Rules [1] [2]

Free User Pack · 1002026-10-20 ↗

USD 20 · one-time payment

Usage / one-time

Usage / one-time 100 general AI credits; total calculated as count × published unit price Rules [1] [2]

Free User Pack · 2002026-10-20 ↗

USD 36 · one-time payment

Usage / one-time

Usage / one-time 200 general AI credits; total calculated as count × published unit price Rules [1] [2]

Free User Pack · 5002026-10-20 ↗

USD 70 · one-time payment

Usage / one-time

Usage / one-time 500 general AI credits; total calculated as count × published unit price Rules [1] [2]

Free User Pack · 10002026-10-20 ↗

USD 120 · one-time payment

Usage / one-time

Usage / one-time 1000 general AI credits; total calculated as count × published unit price Rules [1] [2]

Free User Pack · 20002026-10-20 ↗

USD 200 · one-time payment

Usage / one-time

Usage / one-time 2000 general AI credits; total calculated as count × published unit price Rules [1] [2]

Free User Pack · 50002026-10-20 ↗

USD 400 · one-time payment

Usage / one-time

Usage / one-time 5000 general AI credits; total calculated as count × published unit price Rules [1] [2]

Premium User Pack · 502026-10-20 ↗

USD 6 · one-time payment

Usage / one-time

Usage / one-time 50 general AI credits; total calculated as count × published unit price Rules [1] [2]

Premium User Pack · 1002026-10-20 ↗

USD 10 · one-time payment

Usage / one-time

Usage / one-time 100 general AI credits; total calculated as count × published unit price Rules [1] [2]

Premium User Pack · 2002026-10-20 ↗

USD 18 · one-time payment

Usage / one-time

Usage / one-time 200 general AI credits; total calculated as count × published unit price Rules [1] [2]

Premium User Pack · 5002026-10-20 ↗

USD 35 · one-time payment

Usage / one-time

Usage / one-time 500 general AI credits; total calculated as count × published unit price Rules [1] [2]

Premium User Pack · 10002026-10-20 ↗

USD 60 · one-time payment

Usage / one-time

Usage / one-time 1000 general AI credits; total calculated as count × published unit price Rules [1] [2]

Premium User Pack · 20002026-10-20 ↗

USD 100 · one-time payment

Usage / one-time

Usage / one-time 2000 general AI credits; total calculated as count × published unit price Rules [1] [2]

Premium User Pack · 50002026-10-20 ↗

USD 200 · one-time payment

Usage / one-time

Usage / one-time 5000 general AI credits; total calculated as count × published unit price Rules [1] [2]

19 plan / billing-period records, all listed; — means that period is not recorded.

[1] Premium pricing requires a paid membership; this does not establish removal count

[2] Valid for 12 months after purchase

The feature FAQ confirms watermark-free transparent PNG/blurred-background downloads with a free account; replacing with a new background allows one free download per day, while premium backgrounds require a subscription. The pricing table says Limited without quantifying daily transparent-PNG limits. General AI credits or the one-daily replacement limit are not applied to transparent cutouts.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Cancellation and refund terms follow official checkout and help policies.

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗

Current plans and free limits ↗ · 2026-10-20 Capture 1 ↗

Removal download and free-use rules ↗ · 2026-10-20 Capture 1 ↗

Allowance & disclosure on test date

Free accounts can claim 3 bonus credits daily ('Earn Free Credits -> Claim 3 Credits'); Nano Banana 2 Lite (1:1, 1 image) costs 2 credits per generation

The 'Generate ✦ 2' button explicitly states the 2-credit cost before submission

Evaluation basis & full scores · Overall 87.8/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 78/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	24 / 35	Clicking 'Generate AI Images Now' on the landing page opens the editor workspace; generation requires signing in to a CyberLink account, where the account menu offers 'Earn Free Credits' (3 daily bonus credits). Selecting Nano Banana 2 Lite (1:1, 1 image) clearly shows the 2-credit cost ('Generate ✦ 2') on the submit button and exports an unwatermarked 1024×1024 JPEG.	Evidence A →
Interaction restraint	23 / 30	Requires account sign-in before submitting a prompt, but model, aspect ratio, and image-count controls are clear, and clicking the download icon on the generated image directly saves a 1024×1024 unwatermarked JPEG without secondary upsell modals or watermarks.	Evidence A →
Background & reputation	19 / 20	Developed and operated by CyberLink Corp. (New Taipei City, Taiwan, TWSE: 5203, founded 1996, maker of PowerDirector and PhotoDirector).	Evidence A →
Engineering & privacy	12 / 15	Submits generation via POST https://myedit.online/tti/effect and retrieves the generated result via GET https://myedit.online/tti/effect/{id}/1 (server_s = 7.72s, end_to_end_s = 7.71s), exporting a native 1024×1024 unwatermarked RGB JPEG (191,938 bytes).	Evidence A →

Capability · 92/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	61 / 65	Text rendering (25.0/25: Wooden sign prints two-line 'REALTEST 2026' in clean bold lettering with 100% accuracy) · Count & color binding (16.0/20: Two blue ceramic mugs sit behind the robot on the oak table as requested, while 4 red apples appear in front instead of 3—one extra apple) · Anatomy & structure (20.0/20: Vintage brass robot grips the wooden sign pole with a clearly articulated 5-digit right hand—4 front fingers + thumb on top—and has a complete 5-fingered left hand at its side).	Evidence A →
Deliverable specifications	19 / 20	Clean unwatermarked export (8.0/8); native export is a 1024×1024 (1K) RGB JPEG (191,938 bytes) (11.0/12).	Evidence A →
Processing speed	12 / 15	2026-10-21 same-run CDP network timing: 7.72s server wait from POST /tti/effect (r7) to GET /tti/effect/{id}/1 (r11) (7.71s end-to-end, async_verified) (12.0/15).	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Ideogram

ideogram.ai · Image generation

88/100

Seven-aspect task score

Good text and 2K output come with four candidates per request; free credits depend on sign-in method, exclude email-only signup, and this run took about 34 seconds.

Tested export 2048×2048 (2K Original size unwatermarked JPEG, 785.6KB, 400% pixel retention) · Historical description; output not attached

Historical test · 2026-10-16
Signed-in flow; paid tier not standardized · Fixed English prompt · Historical page record; output not attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Entry & rules ↗Result screen ↗Download & limits ↗Additional screenshot 1 ↗

Use this capability ↗Tool profile →Full report →

Task quality

Text and hand structure performed well on this prompt, with 2048px output; other prompts remain untested.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Text and hand structure performed well on this prompt, with 2048px output; other prompts remain untested.

Original historical observation

Powered by Ideogram 4.0, it is the only platform tested that exports native 2048×2048 (2K) unwatermarked images on the free tier while achieving a 64.0/65 constraint score (#4 overall at 88.2 pts).

Suitable uses · historical assessment

Designers and creators needing native 2048×2048 (2K) unwatermarked exports, world-class typography, and 5-finger anatomical precision

Result & output evidence

No native output obtained; the preview does not establish final export quality.

Open result screenshot ↗

Full test record →

Evidence & scope →

Delivery capabilities

2048×2048 (2K Original size unwatermarked JPEG, 785.6KB, 400% pixel retention) · Historical description; output not attached · Check watermark record

Detailed facts & historical record

Recorded export specifications

2048×2048 (2K Original size unwatermarked JPEG, 785.6KB, 400% pixel retention) · Historical description; output not attached · Check watermark record

No native output is attached; check the evidence for specification and watermark claims.

Historical specification note · not an official maximum

2048×2048 (2K Original Size Unwatermarked JPEG · Ideogram 4.0)

Watermark observation

Zero watermark (clean 2048×2048 2K export)

Batch & API

4-image batch default + Magic Prompt, Remove BG, Upscale, Edit (Ideogram 4.5), and Developer API

Recorded free-tier dimensions

Not verified

Attached file dimensions

Not verified

Evidence & scope →

Speed & reliability

About 34.45s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 33.9s · End-to-end reference: 34.45s
Date & fixture: 2026-10-16 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Must sign in with Google / Apple / Microsoft to receive free credits (Email-only signups receive 0 free credits)
Manual adjustment: Not verified
Interaction observations: Completing Google login (Pass 2) allows skipping the plan upsell ("Skip for now"); default Auto model invokes flagship Ideogram 4.0 (~12 credits), generates 4 concurrent variations, and allows free download of the full 2K (2048×2048) unwatermarked JPEG (03_download.png).

Evidence & scope →

Background & reputation

Ideogram AI, Inc. · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Ideogram AI, Inc.
Recorded origin: Toronto, Ontario, Canada
Domain history record: 2023–present
Background & reputation assessment: Operated by Ideogram AI, Inc. (Toronto, Canada, founded 2023 by former Google Brain diffusion researchers Mohammad Norouzi, William Chan, et al., backed by a16z), globally renowned for AI typography.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Generates 4× 2048×2048 (4.19MP, 400% of 1024² baseline) images per run (~33.9s) and integrates native Remove BG, Upscale, and Ideogram 4.5 Edit apps.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Ideogram official plans (Free / Plus / Pro / Team / Enterprise); USD (before tax) · 2026-10-21

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free (OAuth sign-in)2026-10-21 ↗

Free

Free 30 credits/week when signing in with Google, Apple, or Microsoft (email/password accounts receive no free credits; public generations).

Plus Monthly2026-10-21 ↗

USD 20/month

—

Monthly 2,400 priority credits/month (do not roll over) + unlimited slower queue on Ideogram models + private generation and quality export.

Plus Yearly2026-10-21 ↗

—

USD 180/year

Annual Billed annually at USD 180/year (USD 15/month equivalent); 2,400 priority credits issued monthly + unlimited slower queue + private generation.

Pro Monthly2026-10-21 ↗

USD 60/month

—

Monthly 7,500 priority credits/month (do not roll over) + unlimited slower queue + Batch Generation, larger queue, and private generation.

Pro Yearly2026-10-21 ↗

—

USD 504/year

Annual Billed annually at USD 504/year (USD 42/month equivalent); 7,500 priority credits issued monthly + unlimited slower queue + Batch Generation and private generation.

Team Monthly (per user, min 2)2026-10-21 ↗

USD 30/month

—

Monthly USD 30 per user/month (minimum 2 users); 3,600 priority credits/user/month + unlimited slower queue + central billing and administration.

Team Yearly (per user, min 2)2026-10-21 ↗

—

USD 240/year

Annual USD 240 per user/year (USD 20/user/month equivalent, minimum 2 users); 3,600 priority credits/user/month + unlimited slower queue + central billing.

Enterprise Agreement2026-10-21 ↗

Contact sales

Quoted

Quoted Custom organization agreement and shared Enterprise Prepaid wallet.

8 plan / billing-period records, all listed; — means that period is not recorded.

Official documentation (Plans and Credits) lists Free (30 credits/week when signing in with Google, Apple, or Microsoft; email/password accounts do not get free credits; public generations), Plus (USD 20/month or USD 180/year, i.e., USD 15/month, with 2,400 credits/month + unlimited slow queue + private mode), Pro (USD 60/month or USD 504/year, i.e., USD 42/month, with 7,500 credits/month + batch generation), and Team (min 2 users, USD 30/user/month or USD 240/user/year, 3,600 credits/user/month). Legacy Basic ($8/mo) is no longer sold.

Cancellation, sources & verification record

Monthly plan credits reset on each renewal date and do not roll over; purchased top-up credits are spent after monthly credits. Manage subscriptions and auto-recharge in Settings → Wallet & Billing.

Ideogram official documentation: Plans and Credits pricing table ↗ · 2026-10-21 Capture 1 ↗

Allowance & disclosure on test date

Free credits granted exclusively to Google / Apple / Microsoft OAuth accounts; default Ideogram 4.0 (Auto) costs ~12 credits for 4× 2048×2048 images

Disclosed upfront in the login page footer before registration ("Email accounts don't include free credits")

Evaluation basis & full scores · Overall 87.6/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 89/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	29 / 35	Landing page (01_landing.png) CTA "Get started" leads to the login page (01_pass1_login_modal.png), where bottom fine print explicitly discloses a critical rule: "*Email accounts don't include free credits — you'll need a paid plan to generate. Sign up with Google, Apple, or Microsoft to get free credits."	Evidence A →
Interaction restraint	26 / 30	Completing Google login (Pass 2) allows skipping the plan upsell ("Skip for now"); default Auto model invokes flagship Ideogram 4.0 (~12 credits), generates 4 concurrent variations, and allows free download of the full 2K (2048×2048) unwatermarked JPEG (03_download.png).	Evidence A →
Background & reputation	20 / 20	Operated by Ideogram AI, Inc. (Toronto, Canada, founded 2023 by former Google Brain diffusion researchers Mohammad Norouzi, William Chan, et al., backed by a16z), globally renowned for AI typography.	Evidence C →
Engineering & privacy	14 / 15	Generates 4× 2048×2048 (4.19MP, 400% of 1024² baseline) images per run (~33.9s) and integrates native Remove BG, Upscale, and Ideogram 4.5 Edit apps.	Evidence A →

Capability · 87/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	64 / 65	Text Rendering (25/25: Exact 2-line "REALTEST / 2026" hand-carved typography on the wooden sign with 100% spelling accuracy (25/25)) · Count & Color Binding (19/20: Strictly 3 red apples on the left + 2 stacked blue ceramic mugs on the right (slight twin-handle / 3rd rim artifact behind the lower mug, -1 pt) (19.0/20)) · 5-Finger Hand & Structure (20/20: Vintage brass robotic right hand supporting the bottom of the wooden sign displays 5 crystal-clear, distinct articulated brass fingers (20/20))	Evidence B →
Deliverable specifications	20 / 20	Clean unwatermarked export (8/8); Export & Max deliverable spec: 2048×2048 (2K Original Size Unwatermarked JPEG · Ideogram 4.0)	Evidence B →
Processing speed	3 / 15	Network-observed server wait: 33.9s (API mode: async_poll; end-to-end 34.45s shown for reference only, never penalized)	Evidence B →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Leonardo.Ai

leonardo.ai · Image generation

88/100

Seven-aspect task score

Text and counts perform well with watermark-free exports using free credits; login is required, free results are public, and privacy and deletion require an upgrade.

Tested export 1024×1024 (Unwatermarked JPEG, 147.5KB, 100% pixel retention) · Historical description; output not attached

Historical test · 2026-10-16
Signed-in flow; paid tier not standardized · Fixed English prompt · Historical page record; output not attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Entry & rules ↗Result screen ↗Download & limits ↗Additional screenshot 1 ↗

Use this capability ↗Tool profile →Full report →

Task quality

Text, object counts and colours matched this prompt; other models and prompts require separate checks.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Text, object counts and colours matched this prompt; other models and prompts require separate checks.

Original historical observation

Default Auto model routes to GPT Image 2 and achieves 64.0/65 constraint adherence (#5 overall at 87.6 pts)—just note that free-tier images are public and cannot be deleted without upgrading.

Suitable uses · historical assessment

Creators wanting a generous daily free quota (~18 unwatermarked 1024×1024 images/day via GPT Image 2) and seamless "Edit in Canva" workflow integration

Result & output evidence

No native output obtained; the preview does not establish final export quality.

Open result screenshot ↗

Full test record →

Evidence & scope →

Delivery capabilities

1024×1024 (Unwatermarked JPEG, 147.5KB, 100% pixel retention) · Historical description; output not attached · Check watermark record

Detailed facts & historical record

Recorded export specifications

1024×1024 (Unwatermarked JPEG, 147.5KB, 100% pixel retention) · Historical description; output not attached · Check watermark record

No native output is attached; check the evidence for specification and watermark claims.

Historical specification note · not an official maximum

1024×1024 (Unwatermarked JPEG · Default Auto Routes to GPT Image 2)

Watermark observation

Zero watermark (clean 1024×1024 export)

Batch & API

One-click "Edit in Canva", "Magic Layers", Remove Background, Upscale, Image-to-Video, and Production API

Recorded free-tier dimensions

Not verified

Attached file dimensions

Not verified

Evidence & scope →

Speed & reliability

About 20.24s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 19.8s · End-to-end reference: 20.24s
Date & fixture: 2026-10-16 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Free Canva / Apple / Google / Microsoft / Email login
Manual adjustment: Not verified
Interaction observations: One-click download delivers a clean 1024×1024 unwatermarked JPEG (03_download.png); however, free users are restricted to 1 image per batch (2/3/4 are locked as "Premium option") and cannot delete their own public generations ("Upgrade to Premium to delete image" trash-can lock).

Evidence & scope →

Background & reputation

Leonardo Interactive Pty Ltd (Subsidiary of Canva) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Leonardo Interactive Pty Ltd (Subsidiary of Canva)
Recorded origin: Sydney, Australia (Canva Group)
Domain history record: 2022–present
Background & reputation assessment: Operated by Leonardo Interactive Pty Ltd (Sydney, Australia, founded 2022; acquired by Canva in July 2024), one of the largest native AI creative platforms.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Automatically assigns an anonymous handle (e.g., PixBuilder4344) to protect user email privacy and integrates "Edit in Canva", but penalizes free-tier user control by locking image deletion and Private Mode behind paid plans.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Web individual and default three-seat team plans; excludes API · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free2026-10-20 ↗

Free

Free 150 Fast Tokens/day; public generations and basic quality

Essential2026-10-20 ↗ 2026-10-20 ↗

USD 12/month

USD 120/year

Monthly 8,500 Fast Tokens/month; token bank 25,500

Annual 8,500 Fast Tokens/month Rules [1]

Premium2026-10-20 ↗ 2026-10-20 ↗

USD 30/month

USD 288/year

Monthly 25,000 Fast Tokens/month; token bank 75,000

Annual 25,000 Fast Tokens/month Rules [1]

Ultimate2026-10-20 ↗ 2026-10-20 ↗

USD 60/month

USD 576/year

Monthly 60,000 Fast Tokens/month; token bank 180,000

Annual 60,000 Fast Tokens/month Rules [1]

Starter · 3 seats2026-10-20 ↗

USD 72/month

—

Monthly 75,000 shared tokens; default 3-seat selection

Growth · 3 seats2026-10-20 ↗

USD 144/month

—

Monthly 180,000 shared tokens; default 3-seat selection

Custom2026-10-20 ↗

Contact sales

Quoted

Quoted Custom team and business requirements

10 plan / billing-period records, all listed; — means that period is not recorded.

[1] USD annual price confirmed in official app payment panel; tax and other regional prices depend on checkout

The app payment panel confirms personal annual prices of $120/$288/$576, differing from earlier rounded monthly equivalents on the marketing site. Public plans permit paid users to buy top-ups without quoting pack prices or per-removal token cost. Unlimited relaxed generation does not establish unlimited removal.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Self-serve cancellation in account settings

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗

Official app annual payment selection ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗

Allowance & disclosure on test date

150 free daily tokens reset every day; default Auto (GPT Image 2 · Low) costs 8 tokens per 1024×1024 image (~18 free images/day)

Prompts for free login before generating; Private Mode and Delete buttons display green Premium diamond badges

Evaluation basis & full scores · Overall 87.6/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 82/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	28 / 35	Landing page (01_landing.png) CTA "Generate AI image" leads to app.leonardo.ai and prompts for free sign-in (01_pass1_login_modal.png); free Google login grants 150 daily tokens, and default Auto model (routing to GPT Image 2 · Low) costs only 8 tokens per 1024×1024 generation (~18 free images/day).	Evidence A →
Interaction restraint	24 / 30	One-click download delivers a clean 1024×1024 unwatermarked JPEG (03_download.png); however, free users are restricted to 1 image per batch (2/3/4 are locked as "Premium option") and cannot delete their own public generations ("Upgrade to Premium to delete image" trash-can lock).	Evidence A →
Background & reputation	19 / 20	Operated by Leonardo Interactive Pty Ltd (Sydney, Australia, founded 2022; acquired by Canva in July 2024), one of the largest native AI creative platforms.	Evidence C →
Engineering & privacy	11 / 15	Automatically assigns an anonymous handle (e.g., PixBuilder4344) to protect user email privacy and integrates "Edit in Canva", but penalizes free-tier user control by locking image deletion and Private Mode behind paid plans.	Evidence A →

Capability · 90/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	64 / 65	Text Rendering (25/25: Exact single-line "REALTEST 2026" printed across the dark wooden sign with 100% spelling accuracy (25/25)) · Count & Color Binding (20/20: Strictly 3 red apples on the left + 2 blue ceramic mugs on the right on a dark oak table, zero color bleeding (20/20)) · 5-Finger Hand & Structure (19/20: Vintage brass robotic right hand wraps around the left edge of the wooden sign showing 4 articulated brass fingers in front (thumb supporting from behind) (19.0/20))	Evidence B →
Deliverable specifications	20 / 20	Clean unwatermarked export (8/8); Export & Max deliverable spec: 1024×1024 (Unwatermarked JPEG · Default Auto Routes to GPT Image 2)	Evidence B →
Processing speed	6 / 15	Network-observed server wait: 19.8s (API mode: async_poll; end-to-end 20.24s shown for reference only, never penalized)	Evidence B →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Cutout.pro

cutout.pro · Image generation

87/100

Seven-aspect task score

Clear models and credit costs accompany good text, counts and hands without watermarks; login is required, starter credits cover one image, and stated retention is 30 days.

Tested export 1024×1024 (Unwatermarked JPEG, 667KB) · Historical description; output not attached

Historical test · 2026-10-16
Signed-in flow; paid tier not standardized · Fixed English prompt · Historical page record; output not attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Text, object counts and hand constraints matched this prompt; consistency across other prompts is untested.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Text, object counts and hand constraints matched this prompt; consistency across other prompts is untested.

Original historical observation

Achieved a perfect 65.0/65 constraint quality score thanks to defaulting to GPT Image 2 (1K), sharply outperforming its own 500×500 preview-capped Background Remover (72.9).

Suitable uses · historical assessment

Users who want a free trial of OpenAI's flagship GPT Image 2 model with 100% 5-constraint accuracy (65.0/65 perfect score)

Result & output evidence

No native output obtained; the preview does not establish final export quality.

Open result screenshot ↗

Full test record →

Evidence & scope →

Delivery capabilities

1024×1024 (Unwatermarked JPEG, 667KB) · Historical description; output not attached · Check watermark record

Detailed facts & historical record

Recorded export specifications

1024×1024 (Unwatermarked JPEG, 667KB) · Historical description; output not attached · Check watermark record

No native output is attached; check the evidence for specification and watermark claims.

Historical specification note · not an official maximum

1024×1024 (Unwatermarked JPEG · GPT Image 2)

Watermark observation

Zero watermark (clean 1024×1024 export)

Batch & API

Supports multiple top models (GPT Image 2, Nano Banana, Flux) and commercial API

Recorded free-tier dimensions

Not verified

Attached file dimensions

Not verified

Evidence & scope →

Speed & reliability

About 26.2s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 24.5s · End-to-end reference: 26.2s
Date & fixture: 2026-10-16 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Free Google / Facebook / Email login (grants 5 initial credits)
Manual adjustment: Not verified
Interaction observations: Clean workspace without deceptive fake progress bars or watermark traps; upfront credit cost (3 credits) is clearly labeled on the CTA button.

Evidence & scope →

Background & reputation

LibAI Tech Limited (Cutout.Pro) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: LibAI Tech Limited (Cutout.Pro)
Recorded origin: Hong Kong
Domain history record: 2018–present
Background & reputation assessment: Operated by LibAI Tech Ltd. (Cutout.Pro, Hong Kong, founded 2018), a specialized visual AI platform.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Calls GPT Image 2 backend asynchronously (24.5s server inference) and delivers a direct 1024×1024 unwatermarked download.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Web image removal: free, all expanded monthly tiers and packs; API priced separately · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free account2026-10-20 ↗

Free

Free Free previews; 5 signup credits

80 credits2026-10-20 ↗

USD 5/month

—

Monthly 80 credits/month; removal 1 credit/image Rules [1]

170 credits2026-10-20 ↗

USD 9.9/month

—

Monthly 170 credits/month; removal 1 credit/image Rules [1]

300 credits2026-10-20 ↗

USD 16.9/month

—

Monthly 300 credits/month; removal 1 credit/image Rules [1]

550 credits2026-10-20 ↗

USD 29/month

—

Monthly 550 credits/month; removal 1 credit/image Rules [1]

1,000 credits2026-10-20 ↗

USD 49/month

—

Monthly 1,000 credits/month; removal 1 credit/image Rules [1]

1,450 credits2026-10-20 ↗

USD 69/month

—

Monthly 1,450 credits/month; removal 1 credit/image Rules [1]

2,400 credits2026-10-20 ↗

USD 108/month

—

Monthly 2,400 credits/month; removal 1 credit/image Rules [1]

5,600 credits2026-10-20 ↗

USD 239/month

—

Monthly 5,600 credits/month; removal 1 credit/image Rules [1]

10,000 credits2026-10-20 ↗

USD 399/month

—

Monthly 10,000 credits/month; removal 1 credit/image Rules [1]

15,000 credits2026-10-20 ↗

USD 549/month

—

Monthly 15,000 credits/month; removal 1 credit/image Rules [1]

25,000 credits2026-10-20 ↗

USD 859/month

—

Monthly 25,000 credits/month; removal 1 credit/image Rules [1]

50,000 credits2026-10-20 ↗

USD 1,599/month

—

Monthly 50,000 credits/month; removal 1 credit/image Rules [1]

75,000 credits2026-10-20 ↗

USD 2,299/month

—

Monthly 75,000 credits/month; removal 1 credit/image Rules [1]

150,000 credits2026-10-20 ↗

USD 4,049/month

—

Monthly 150,000 credits/month; removal 1 credit/image Rules [1]

250,000 credits2026-10-20 ↗

USD 5,749/month

—

Monthly 250,000 credits/month; removal 1 credit/image Rules [1]

1,000,000 credits2026-10-20 ↗

USD 20,000/month

—

Monthly 1,000,000 credits/month; removal 1 credit/image Rules [1]

2,000,000 credits2026-10-20 ↗

USD 34,000/month

—

Monthly 2,000,000 credits/month; removal 1 credit/image Rules [1]

30 credits2026-10-20 ↗

USD 4.99 · one-time payment

Usage / one-time

Usage / one-time 30 credits; removal 1 credit/image Rules [2]

300 credits2026-10-20 ↗

USD 39 · one-time payment

Usage / one-time

Usage / one-time 300 credits; removal 1 credit/image Rules [2]

900 credits2026-10-20 ↗

USD 79 · one-time payment

Usage / one-time

Usage / one-time 900 credits; removal 1 credit/image Rules [2]

3,000 credits2026-10-20 ↗

USD 199 · one-time payment

Usage / one-time

Usage / one-time 3,000 credits; removal 1 credit/image Rules [2]

30,000 credits2026-10-20 ↗

USD 1,399 · one-time payment

Usage / one-time

Usage / one-time 30,000 credits; removal 1 credit/image Rules [2]

5,000,000 credits2026-10-20 ↗

USD 37,500 · one-time payment

Usage / one-time

Usage / one-time 5,000,000 credits; removal 1 credit/image Rules [2]

Custom2026-10-20 ↗

Contact sales

Quoted

Quoted Custom subscription or credit pack

25 plan / billing-period records, all listed; — means that period is not recorded.

[1] Rollover capped at 5× monthly allowance

[2] No expiry

High-volume tiers expanded; removal costs 1 credit/image, other operations differ. Taxes are not stated on the cards.

Cancellation, sources & verification record

Downgrade, upgrade or cancel anytime; advertised 14-day money-back guarantee subject to terms.

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗

Allowance & disclosure on test date

5 free credits on new account signup; default GPT Image 2 (1K) model costs 3 credits per image (1 free generation)

Upfront credit cost ("Generate (3)") shown directly on the primary button

Evaluation basis & full scores · Overall 87.4/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 79/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	25 / 35	Landing page (01_landing.png) explicitly displays the default model "Gpt Image 2 · 1K · 1:1" and button "Generate (3)" (3 credits); free Google registration grants 5 free credits, enough to generate and download 1 full 1024×1024 unwatermarked image.	Evidence A →
Interaction restraint	24 / 30	Clean workspace without deceptive fake progress bars or watermark traps; upfront credit cost (3 credits) is clearly labeled on the CTA button.	Evidence A →
Background & reputation	17 / 20	Operated by LibAI Tech Ltd. (Cutout.Pro, Hong Kong, founded 2018), a specialized visual AI platform.	Evidence C →
Engineering & privacy	13 / 15	Calls GPT Image 2 backend asynchronously (24.5s server inference) and delivers a direct 1024×1024 unwatermarked download.	Evidence A →

Capability · 91/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	65 / 65	Text Rendering (25/25: Flawless single-line stencil "REALTEST 2026" centered on the wooden sign with natural woodgrain texture (25/25)) · Count & Color Binding (20/20: Strictly 3 red apples on the left + 2 dark-blue ceramic mugs on the right on a dark oak table (20/20)) · 5-Finger Hand & Structure (20/20: Both brass robotic hands exhibit 4 articulated fingers + 1 thumb = 5 distinct mechanical digits with realistic metallic patina (20/20))	Evidence B →
Deliverable specifications	20 / 20	Clean unwatermarked export (8/8); Export & Max deliverable spec: 1024×1024 (Unwatermarked JPEG · GPT Image 2)	Evidence B →
Processing speed	6 / 15	Network-observed server wait: 24.5s (API mode: async_poll; end-to-end 26.2s shown for reference only, never penalized)	Evidence B →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Img2Go

img2go.com · Image generation

87/100

Seven-aspect task score

Free login-free generation returns four clean images without blocking popups or a public guest gallery; object counts and hands still missed parts of the prompt.

Tested export 1024 × 1024 · PNG

Historical test · 2026-10-19
Guest flow · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (1024 × 1024 · PNG) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Generated four images without login with correct text in the reviewed result; findings cover this prompt only.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Generated four images without login with correct text in the reviewed result; findings cover this prompt only.

Original historical observation

Img2Go AI Creator Studio generates four 1024×1024 candidates simultaneously in 14.0s without login, nailing the exact 'REALTEST 2026' wooden sign lettering and warm coffee-shop lighting.

Suitable uses · historical assessment

Users needing fast, no-login 4-up 1024×1024 image generation with reliable English sign typography

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

1024 × 1024 · PNG · No watermark in test

Detailed facts & historical record

Recorded export specifications: 1024 × 1024 · PNG · No watermark in test
Historical specification note · not an official maximum: 1024×1024 (Generates 4 unwatermarked 1024×1024 images per prompt)
Watermark observation: Watermark-Free
Batch & API: Generates 4 variations per prompt in a single run
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: 1024 × 1024 · PNG

Evidence & scope →

Speed & reliability

About 14.2s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 14s · End-to-end reference: 14.2s
Date & fixture: 2026-10-19 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

No login in this run

Detailed facts & historical record

Account requirements: No login in this test
Sign-in methods: No login required
Manual adjustment: Not verified
Interaction observations: Clean studio workspace with aspect ratio (Square 1:1 / Portrait / Landscape) and style selectors, zero forced signup popups before download.

Evidence & scope →

Background & reputation

QaamGo Web GmbH (Radolfzell, Germany) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: QaamGo Web GmbH (Radolfzell, Germany)
Recorded origin: Radolfzell, Germany
Domain history record: 2016–Present
Background & reputation assessment: Operated by Germany's QaamGo Web GmbH (Radolfzell).
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Claims deletion within 1 hours; deletion not independently verified

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: EU GDPR compliant cloud generation workflow with no personal account required for guest runs.
Retention statement: Claims deletion within 1 hours; deletion not independently verified

Evidence & scope →

Pricing & rules

Website free, monthly, annual and education plans; USD, Business billed per person · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free2026-10-20 ↗

Free

Free Limited processing; no universal free-credit count published

Starter2026-10-20 ↗

USD 8/month

USD 77/year

Monthly Baseline 1× AI usage; batch processing

Annual Baseline 1× AI usage; 64GB files and batch processing

Pro2026-10-20 ↗

USD 29/month

USD 278/year

Monthly 5× Starter AI usage

Annual 5× Starter AI usage

Business2026-10-20 ↗

USD 89/month

USD 854/year

Monthly Per-user billing; 20× Starter AI usage

Annual Per-user billing; 20× Starter AI usage

Education program2026-10-20 ↗

Free

Free Paid apps free after student/teacher eligibility verification; non-commercial educational use only

8 plan / billing-period records, all listed; — means that period is not recorded.

Actual annual amounts are $77/$278/$854. Absolute removal quotas and per-image credit usage are unpublished; 1×/5×/20× cannot be converted into image counts.

Cancellation, sources & verification record

Page advertises a 14-day money-back guarantee and cancellation anytime; eligibility is subject to terms.

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗

Allowance & disclosure on test date

Free visitors can generate a 4-image 1024×1024 batch and download directly without signup

Clearly displays aspect ratio and style options prior to generation

Evaluation basis & full scores · Overall 86.9/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 93.5/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	33.5 / 35	Clicking 'Start Generating' on the landing page (01_landing.png) opens Img2Go's AI Creator Studio, which generates 4 full 1024×1024 images in one batch and allows direct unwatermarked download without login (02_result.png, 03_download.png).	Evidence A →
Interaction restraint	27 / 30	Clean studio workspace with aspect ratio (Square 1:1 / Portrait / Landscape) and style selectors, zero forced signup popups before download.	Evidence A →
Background & reputation	19 / 20	Operated by Germany's QaamGo Web GmbH (Radolfzell).	Evidence A →
Engineering & privacy	14 / 15	EU GDPR compliant cloud generation workflow with no personal account required for guest runs.	Evidence A →

Capability · 84/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	54 / 65	Text Rendering (25/25: exact 'REALTEST 2026' on the wooden sign) · Count & Color Binding (16/20: 2 blue mugs on left + 4 red apples on right instead of 3) · 5-Finger Hand & Structure (13/20: barista hand mostly occluded behind the portafilter and wooden sign supported by two small counter clamps) (54.0/65)	Evidence A →
Deliverable specifications	18.5 / 20	Generates 4 unwatermarked 1024×1024 JPEG variations per run without login (20.0/20)	Evidence A →
Processing speed	11.5 / 15	Generates 4× 1024×1024 images in ~14.0s (11.5/15)	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Pixelcut

pixelcut.ai · Image generation

87/100

Seven-aspect task score

Switching to the free model yields four 1600px clean images with good text and counts; login is required and the default Pro model first triggers a paywall.

Tested export 1600×1600 (Unwatermarked PNG/JPEG) · Historical description; output not attached

Historical test · 2026-10-16
Signed-in flow; paid tier not standardized · Fixed English prompt · Historical page record; output not attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Text, counts and hand structure matched this prompt; consistency in other scenes was not retested.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Text, counts and hand structure matched this prompt; consistency in other scenes was not retested.

Original historical observation

Near-perfect 5-point constraint adherence (64.5/65): exact "REALTEST 2026" text, exact 3 red apples + 2 blue ceramic mugs, and a 5-finger articulated brass hand.

Suitable uses · historical assessment

Users needing exact English typography rendering, strict multi-object count/color adherence, and 1600×1600 unwatermarked exports

Result & output evidence

No native output obtained; the preview does not establish final export quality.

Open result screenshot ↗

Full test record →

Evidence & scope →

Delivery capabilities

1600×1600 (Unwatermarked PNG/JPEG) · Historical description; output not attached · Check watermark record

Detailed facts & historical record

Recorded export specifications

1600×1600 (Unwatermarked PNG/JPEG) · Historical description; output not attached · Check watermark record

No native output is attached; check the evidence for specification and watermark claims.

Historical specification note · not an official maximum

1600×1600 (Unwatermarked Full-Res)

Watermark observation

Zero watermark (clean 1600×1600 export)

Batch & API

Generates 4 variations per prompt with one-click Background Removal / Upscale / Retouch integration

Recorded free-tier dimensions

Not verified

Attached file dimensions

Not verified

Evidence & scope →

Speed & reliability

About 21.2s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 19.4s · End-to-end reference: 21.2s
Date & fixture: 2026-10-16 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Free Google / Apple / Email login (grants 108 initial credits)
Manual adjustment: Not verified
Interaction observations: Google sign-in preserves the exact prompt typed on the landing page, though defaulting to a Pro-locked model creates an unnecessary paywall detour before discovering the free model.

Evidence & scope →

Background & reputation

Pixelcut, Inc. (YC S20) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Pixelcut, Inc. (YC S20)
Recorded origin: San Francisco, California, USA
Domain history record: 2020–present
Background & reputation assessment: Operated by Pixelcut, Inc. (Y Combinator S20, San Francisco, CA), offering a multi-model visual workspace.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Free Muse Image model generates 4 concurrent 1600×1600 unwatermarked images (assets.pixelcut.ai) in ~19.4s Network-observed server wait.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Web Free, Pro and all eight Business volumes with monthly/annual selection; API separate · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free2026-10-20 ↗

Free

Free Limited background removal; watermark-free export, exact count unspecified

Pro2026-10-20 ↗

USD 10/month

USD 96/year

Monthly Unlimited removal; 600 AI credits/month; 3-person team; 1,000 batch exports/month

Annual Unlimited removal; 600 AI credits/month; 3-person team; 1,000 batch exports/month Rules [1]

Business 3,6002026-10-20 ↗

USD 30/month

USD 288/year

Monthly Unlimited removal; 3,600 AI credits/month; 10-person team; 2,000 batch exports/month

Annual Unlimited removal; 3,600 AI credits/month; 10-person team; 2,000 batch exports/month Rules [1]

Business 9,0002026-10-20 ↗

USD 60/month

USD 576/year

Monthly Unlimited removal; 9,000 AI credits/month; 20-person team; 5,000 batch exports/month

Annual Unlimited removal; 9,000 AI credits/month; 20-person team; 5,000 batch exports/month Rules [1]

Business 18,0002026-10-20 ↗

USD 120/month

USD 1,152/year

Monthly Unlimited removal; 18,000 AI credits/month; 20-person team; 10,000 batch exports/month

Annual Unlimited removal; 18,000 AI credits/month; 20-person team; 10,000 batch exports/month Rules [1]

Business 36,0002026-10-20 ↗

USD 240/month

USD 2,304/year

Monthly Unlimited removal; 36,000 AI credits/month; 20-person team; 20,000 batch exports/month

Annual Unlimited removal; 36,000 AI credits/month; 20-person team; 20,000 batch exports/month Rules [1]

Business 72,0002026-10-20 ↗

USD 480/month

USD 4,608/year

Monthly Unlimited removal; 72,000 AI credits/month; 20-person team; 40,000 batch exports/month

Annual Unlimited removal; 72,000 AI credits/month; 20-person team; 40,000 batch exports/month Rules [1]

Business 108,0002026-10-20 ↗

USD 720/month

USD 6,912/year

Monthly Unlimited removal; 108,000 AI credits/month; 20-person team; 40,000 batch exports/month

Annual Unlimited removal; 108,000 AI credits/month; 20-person team; 40,000 batch exports/month Rules [1]

Business 144,0002026-10-20 ↗

USD 960/month

USD 9,216/year

Monthly Unlimited removal; 144,000 AI credits/month; 20-person team; 60,000 batch exports/month

Annual Unlimited removal; 144,000 AI credits/month; 20-person team; 60,000 batch exports/month Rules [1]

Business 180,0002026-10-20 ↗

USD 1,200/month

USD 11,520/year

Monthly Unlimited removal; 180,000 AI credits/month; 20-person team; 80,000 batch exports/month

Annual Unlimited removal; 180,000 AI credits/month; 20-person team; 80,000 batch exports/month Rules [1]

19 plan / billing-period records, all listed; — means that period is not recorded.

[1] One annual payment; USD selected at checkout. Bank exchange fees and final invoice conditions must be checked before payment.

Annual charges for Pro and all eight Business volumes were verified separately in Stripe checkout. Unlimited paid removal coexists with batch-export limits. The free tier still says Limited without an exact removal count; it is not treated as unlimited.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Cancel anytime; refund eligibility unverified.

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗

Annual USD checkout summaries for each volume ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗ · Capture 4 ↗ · Capture 5 ↗ · Capture 6 ↗ · Capture 7 ↗ · Capture 8 ↗ · Capture 9 ↗

Allowance & disclosure on test date

108 free credits upon sign-in; switching to free Muse Image (Meta) model costs only 12 credits per 4-image batch

Prompts for free login on Generate; default model displays a Pro padlock icon

Evaluation basis & full scores · Overall 86.8/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 78/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	23 / 35	Landing page (01_landing.png) states "Free to try · No watermarks"; clicking Generate prompts for free Google login (granting 108 credits), but defaults to the Pro-locked model "Nano Banana 2 🔒" which triggers an upgrade paywall unless the user manually switches the model selector to the free "Muse Image (Meta)" model (12 credits per 4-image batch).	Evidence A →
Interaction restraint	23 / 30	Google sign-in preserves the exact prompt typed on the landing page, though defaulting to a Pro-locked model creates an unnecessary paywall detour before discovering the free model.	Evidence A →
Background & reputation	18 / 20	Operated by Pixelcut, Inc. (Y Combinator S20, San Francisco, CA), offering a multi-model visual workspace.	Evidence C →
Engineering & privacy	14 / 15	Free Muse Image model generates 4 concurrent 1600×1600 unwatermarked images (assets.pixelcut.ai) in ~19.4s Network-observed server wait.	Evidence A →

Capability · 90.5/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	64.5 / 65	Text Rendering (25/25: Exact 2-line stencil "REALTEST 2026" on the wooden sign with 100% spelling accuracy and woodgrain perspective (25/25)) · Count & Color Binding (20/20: Strictly 3 red apples with water droplets on the left + 2 speckled blue ceramic mugs on the right, zero color bleeding (20/20)) · 5-Finger Hand & Structure (19.5/20: Vintage brass robotic right hand displays 4 fingers + 1 thumb (5 distinct articulated digits) gripping the wooden sign (19.5/20))	Evidence B →
Deliverable specifications	20 / 20	Clean unwatermarked export (8/8); Export & Max deliverable spec: 1600×1600 (Unwatermarked Full-Res)	Evidence B →
Processing speed	6 / 15	Network-observed server wait: 19.4s (API mode: async_poll; end-to-end 21.2s shown for reference only, never penalized)	Evidence B →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Photoroom

photoroom.com · Image generation

86/100

Seven-aspect task score

Workspace 'Create any image' generates two unwatermarked 1024×1024 JPEGs in 5.98s server wait; sign text and 5-finger mechanical hand are accurate, missing one blue mug.

Tested export JPEG · Output attached

Historical test · 2026-10-21
Guest flow · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (JPEG · Output attached) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

'REALTEST 2026' lettering and 5-finger brass hand are accurate; all 3 red apples appear, but only 1 of 2 blue mugs is rendered (15/20).

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

'REALTEST 2026' lettering and 5-finger brass hand are accurate; all 3 red apples appear, but only 1 of 2 blue mugs is rendered (15/20).

Original historical observation

Strong multi-constraint adherence (59.0/65): 'REALTEST 2026' text and 5-finger mechanical right hand are crisp and accurate; misses one of the two requested blue ceramic mugs (3 red apples + 1 blue mug).

Suitable uses · historical assessment

Users who need fast, accurate text-on-sign rendering and clean 1024×1024 square visuals inside the Photoroom workspace

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

JPEG · Output attached · Check watermark record

Detailed facts & historical record

Recorded export specifications: JPEG · Output attached · Check watermark record
Historical specification note · not an official maximum: 1024×1024 (JPEG)
Watermark observation: No visible watermark (clean 1024×1024 JPEG export)
Batch & API: Generates 2 candidates per run; supports generating 2 variations or opening in chat editor
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: Not verified

Evidence & scope →

Speed & reliability

About 6.54s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 5.98s · End-to-end reference: 6.54s
Date & fixture: 2026-10-21 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

No login in this run

Detailed facts & historical record

Account requirements: No login in this test
Sign-in methods: Accessible via app.photoroom.com AI mini-apps dialog (supports Guest or Google/Apple/Email workspace)
Manual adjustment: Not verified
Interaction observations: Requires opening the AI mini-apps modal and selecting 'Create any image', then downloading via the '...' More actions menu; no forced trial or paywall blocks standard 1024×1024 JPEG export.

Evidence & scope →

Background & reputation

PhotoRoom SAS (Paris, France) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: PhotoRoom SAS (Paris, France)
Recorded origin: Paris, France
Domain history record: 2019–present
Background & reputation assessment: Operated by PhotoRoom SAS (Paris, France, founded 2019), a top-tier visual AI company.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Streams generation events over HTTPS via ai-gateway.photoroom.com/v2/ai-tools/generate-images and delivers 1024×1024 JPEGs from cdn.photoroom.com (network-observed server wait 5.98s).
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

USD web subscriptions: Pro, Max and all four Ultra volumes, monthly and annual; excludes API/MCP · 2026-10-21

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free2026-10-20 ↗

Free

Free Limited AI credits and exports; exact quantities are not stated in official help

Personal, non-commercial use only; no API access

Pro2026-10-20 ↗

USD 12.99/month

USD 89.99/year

Monthly 8,000 AI credits/month, 1,000 exports/month; core removal uses no AI credits

Annual 8,000 AI credits/month, 1,000 exports/month; core removal uses no AI credits

Public card shows $7.50/month equivalent; in-app upgrade modal after sign-in confirms USD 89.99/year (one free trial per person)

Max2026-10-20 ↗

USD 34.99/month

USD 249.99/year

Monthly 25,000 AI credits/month, 3,000 exports/month; core removal uses no AI credits

Annual 25,000 AI credits/month, 3,000 exports/month; core removal uses no AI credits

Public card shows $20.99/month equivalent; in-app upgrade modal after sign-in confirms USD 249.99/year ($20.83/month equivalent, one free trial per person)

Ultra 75k2026-10-20 ↗

USD 99/month

USD 990/year

Monthly 75,000 AI credits/month, 10,000 exports/month; core removal uses no AI credits

Annual 75,000 AI credits/month, 10,000 exports/month; core removal uses no AI credits

Public card shows $82.50/month equivalent; in-app upgrade modal after sign-in confirms USD 990.00/year

Ultra 150k2026-10-20 ↗

USD 198/month

USD 1,980/year

Monthly 150,000 AI credits/month, 20,000 exports/month; core removal uses no AI credits

Annual 150,000 AI credits/month, 20,000 exports/month; core removal uses no AI credits

Public card shows $165/month equivalent; in-app upgrade modal after sign-in confirms USD 1,980.00/year

Ultra 375k2026-10-20 ↗

USD 495/month

USD 4,950/year

Monthly 375,000 AI credits/month, 50,000 exports/month; core removal uses no AI credits

Annual 375,000 AI credits/month, 50,000 exports/month; core removal uses no AI credits

Public card shows $413/month equivalent ($412.50/month in modal); in-app upgrade modal after sign-in confirms USD 4,950.00/year

Ultra 750k2026-10-20 ↗

USD 990/month

USD 9,900/year

Monthly 750,000 AI credits/month, 100,000 exports/month; core removal uses no AI credits

Annual 750,000 AI credits/month, 100,000 exports/month; core removal uses no AI credits

Public card shows $825/month equivalent; in-app upgrade modal after sign-in confirms USD 9,900.00/year

13 plan / billing-period records, all listed; — means that period is not recorded.

Free use is personal and non-commercial, with no exact allowance stated. Core removal uses no AI credits but export caps and fair-use limits apply. The public pricing page only shows monthly equivalents for annual plans; after sign-in, the in-app upgrade modal confirms all six annual invoice totals: Pro USD 89.99/year (not $7.5×12=$90), Max USD 249.99/year ($20.83/month equivalent, not $20.99×12), and Ultra tiers at USD 990 / 1,980 / 4,950 / 9,900/year. The pricing page states VAT/GST is included.

Cancellation, sources & verification record

Cancel in app. Pro/Max annual plans have a seven-day trial. Pro/Max/Ultra refunds may be requested within seven days of the last payment, but are not guaranteed and may be refused for misuse or terms violations. API is excluded.

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗ · Capture 4 ↗ · Capture 5 ↗ · Capture 6 ↗ · Capture 7 ↗ · Capture 8 ↗ · Capture 9 ↗ · Capture 10 ↗ · Capture 11 ↗

Purchase entry requires sign-in ↗ · 2026-10-20 Capture 1 ↗

Official plans and free-use scope ↗ · 2026-10-20 Capture 1 ↗

Official refund eligibility ↗ · 2026-10-20 Capture 1 ↗

In-app upgrade modal annual billing totals (verified after sign-in) ↗ · 2026-10-21 Capture 1 ↗

Allowance & disclosure on test date

Generates 2 square 1024×1024 images per prompt in 'Create any image' with direct JPEG download via More actions

Landing page links directly to the AI mini-apps dialog; Pro/Max upgrade tiers are optional

Evaluation basis & full scores · Overall 86.2/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 82/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	26 / 35	Clicking Generate images on /tools/ai-images opens the app.photoroom.com AI mini-apps dialog where 'Create any image' generates two 1024×1024 square candidates and downloads a clean 1024×1024 JPEG without watermarks or payment prompts (Pro $89.99/yr and Max $249.99/yr verified in workspace billing).	Evidence A →
Interaction restraint	24 / 30	Requires opening the AI mini-apps modal and selecting 'Create any image', then downloading via the '...' More actions menu; no forced trial or paywall blocks standard 1024×1024 JPEG export.	Evidence A →
Background & reputation	19 / 20	Operated by PhotoRoom SAS (Paris, France, founded 2019), a top-tier visual AI company.	Evidence B →
Engineering & privacy	13 / 15	Streams generation events over HTTPS via ai-gateway.photoroom.com/v2/ai-tools/generate-images and delivers 1024×1024 JPEGs from cdn.photoroom.com (network-observed server wait 5.98s).	Evidence A →

Capability · 88/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	59 / 65	Text rendering (25.0/25: Wooden sign prints 'REALTEST 2026' with 100% accurate spelling and metallic embossing) · Count & color binding (15.0/20: 3 red apples on the left are exact, but only 1 blue ceramic mug appears in the background instead of 2) · Anatomy & structure (19.0/20: Vintage brass robot right hand clearly shows 4 mechanical fingers + 1 thumb holding the lower-right corner of the sign; upper torso partially cropped).	Evidence A →
Deliverable specifications	17 / 20	Clean unwatermarked export (8/8); tested native export is 1024×1024 RGB JPEG (245,931 bytes) matching the 1:1 target resolution (10/12).	Evidence A →
Processing speed	12 / 15	Single-sample network-observed server wait 5.98s (fast 5.0–10.0s, async_verified SSE stream from POST /v2/ai-tools/generate-images to CDN image delivery; end-to-end 6.54s) (12.0/15).	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

creator.nightcafe.studio

creator.nightcafe.studio · Image generation

86/100

Seven-aspect task score

NightCafe (Flux 2 Klein 9B Fast) excelled in multi-constraint benchmarks: exact 'REALTEST 2026' text rendering, strict binding of 3 red apples and 2 blue mugs, fast no-login trial, with a small watermark on free exports.

Tested export PNG · Output attached

Historical test · 2026-11-01
Guest flow · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

creator.nightcafe.studio native output thumbnail

Inspect native image (PNG · Output attached) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Superb text spelling and multi-object count/color binding; realistic brass robot aesthetics, full 1024px native resolution with a corner watermark.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Superb text spelling and multi-object count/color binding; realistic brass robot aesthetics, full 1024px native resolution with a corner watermark.

Original historical observation

NightCafe (Flux 2 Klein 9B Fast) demonstrates top-tier semantic following in multi-constraint benchmark: 100% text accuracy, zero count/color bleeding, with impressive metallic and ceramic lighting.

Suitable uses · historical assessment

Creators seeking high-quality multi-model AI generation with top text spelling precision

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

PNG · Output attached · Check watermark record

Detailed facts & historical record

Recorded export specifications: PNG · Output attached · Check watermark record
Historical specification note · not an official maximum: 1024 x 1024 (Full-resolution watermarked JPEG)
Watermark observation: Subtle translucent brand watermark in corner
Batch & API: Multi-model switching, community challenges, private collection controls
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: Not verified

Evidence & scope →

Speed & reliability

About 11s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 10.5s · End-to-end reference: 11s
Date & fixture: 2026-11-01 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

No login in this run

Detailed facts & historical record

Account requirements: No login in this test
Sign-in methods: No-login trial available; Google, Apple, Facebook SSO and email supported
Manual adjustment: Not verified
Interaction observations: Not verified

Evidence & scope →

Background & reputation

NightCafe Studio Pty Ltd · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: NightCafe Studio Pty Ltd
Recorded origin: Australia (Cairns)
Domain history record: Established in 2019
Background & reputation assessment: Not verified
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Google Tag Manager · Stripe · PayPal · Segment
Engineering & privacy assessment: Not verified
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

NightCafe Creator fast credits and PRO membership tiers · 2026-11-01

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free Tier2026-11-01 ↗

Free

Free 5 free credits daily upon login, plus credits earned via daily community challenges

Access to standard image generation models and base upscaling；Standard queues; advanced models consume credits faster; ad-supported；Refreshes daily, no renewal；Perpetual access

AI Hobbyist Monthly2026-11-01 ↗

USD 12.5/month

—

Monthly 200 Fast Credits/month (unexpiring), plus 600+ monthly relax credits

Access to all image models, unlimited Flux 2 Klein / Z-Image Turbo, 5 characters, ad-free；200 Fast Credits per month；Auto-renews monthly; pause or cancel anytime

Rules [1]

AI Hobbyist Annual2026-11-01 ↗

—

USD 89.91/year

Annual 200 Fast Credits/month delivered upfront, billed as $89.91/year ($7.49/mo equivalent)

Full Hobbyist features with 40% annual savings；Prepaid annually at $89.91

Rules [2] [3]

AI Enthusiast Monthly2026-11-01 ↗

USD 25/month

—

Monthly 500 Fast Credits/month, plus 600+ relax credits

Access to all video and image models, unlimited HiDream, 10 characters, private search；500 Fast Credits per month；Auto-renews monthly

Rules [1]

AI Enthusiast Annual2026-11-01 ↗

—

USD 179.91/year

Annual 500 Fast Credits/month, billed as $179.91/year ($14.99/mo equivalent)

Full Enthusiast features with 40% annual savings；Prepaid annually at $179.91

Rules [2] [3]

5 plan / billing-period records, all listed; — means that period is not recorded.

[1] Active during billing month

[2] Auto-renews annually

[3] Active during subscription year

Offers daily free credits and monthly/annual PRO subscriptions; annual plans receive a 40% discount with exact 12-month billed totals explicitly printed.

Cancellation, sources & verification record

Cancellation and refund terms unverified

NightCafe PRO Official Subscription Pricing ↗ · 2026-11-01 Capture 1 ↗

Allowance & disclosure on test date

Free trial without login; 5 free credits daily for registered users

Watermark and export dimensions clearly disclosed in result view

Evaluation basis & full scores · Overall 86.1/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 88.5/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	31 / 35	Explanation unavailable	Evidence A →
Interaction restraint	26 / 30	Explanation unavailable	Evidence A →
Background & reputation	18 / 20	Explanation unavailable	Evidence A →
Engineering & privacy	13.5 / 15	Explanation unavailable	Evidence A →

Capability · 85/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	64 / 65	Text Rendering (25/25: Sign text 'REALTEST 2026' flawless with natural wood grain) · Count & Color Binding (20/20: Strictly 3 red apples and 2 blue ceramic mugs with zero color bleeding) · 5-Finger Hand & Structure (19/20: Articulated brass robotic fingers with excellent metallic reflections)	Evidence A →
Deliverable specifications	12 / 20	Clean unwatermarked export (8/8); Export & Max deliverable spec: 1024 x 1024 (Full-resolution watermarked JPEG)	Evidence A →
Processing speed	9 / 15	Network-observed server wait: 10.5s (API mode: async_poll; end-to-end 11s shown for reference only, never penalized)	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Freepik

freepik.com · Image generation

86/100

Seven-aspect task score

Free login generates a 4-image batch and exports 1024px watermark-free PNGs with fast inference and accurate sign text; requires manually disabling default 16:9 and prompt rewriting, and missed one apple.

Tested export 1024 × 1024 · PNG

Historical test · 2026-10-19
Signed-in flow; paid tier not standardized · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (1024 × 1024 · PNG) ↗ Entry & rules ↗Result screen ↗

The former Freepik feature and pricing pages currently redirect to Magnific; historical test identity and date are preserved. Official notice ↗

Use this capability ↗Tool profile →Full report →

Task quality

Sign text is 100% accurate with detailed brass texture; however, only 2 red apples appeared instead of 3, and the robotic thumb was obscured.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Sign text is 100% accurate with detailed brass texture; however, only 2 red apples appeared instead of 3, and the robotic thumb was obscured.

Original historical observation

Under canonical square settings (explicit 1:1, AI prompt OFF), Freepik / Magnific’s Auto model renders a 4-image 1024×1024 batch in 5.76s: the "REALTEST 2026" sign text is 100% accurate (25/25) with rich brass texture, though it generates 2 red apples instead of 3 and obscures the thumb behind the sign (56.0/65).

Suitable uses · historical assessment

Creators who want a fast (~5.8s) 4-image batch of unwatermarked 1024×1024 PNGs and accurate sign text rendering with a free account

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

1024 × 1024 · PNG · No watermark in test

Detailed facts & historical record

Recorded export specifications: 1024 × 1024 · PNG · No watermark in test
Historical specification note · not an official maximum: 1024×1024 (Unwatermarked Native PNG · Auto Model 4-Image Batch)
Watermark observation: Zero watermark (clean 1024×1024 PNG export)
Batch & API: Generates 4 candidates per batch by default, with in-workspace handoff to Upscale, Relight, and Mystic models
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: 1024 × 1024 · PNG

Evidence & scope →

Speed & reliability

About 6.12s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 5.76s · End-to-end reference: 6.12s
Date & fixture: 2026-10-19 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Free Google / Apple / Email login (automatically resumes pending generation after sign-in)
Manual adjustment: Not verified
Interaction observations: Free download delivers a clean 1024×1024 native PNG without watermarks or credit-card walls; however, the workspace defaults to 16:9 aspect ratio with "AI prompt" rewriting enabled, requiring manual adjustment to 1:1 and turning "AI prompt" OFF for faithful square prompt adherence.

Evidence & scope →

Background & reputation

Freepik Company S.L. / Magnific AI · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: The former Freepik feature and pricing pages currently redirect to Magnific; historical test identity and date are preserved.
Operator: Freepik Company S.L. / Magnific AI
Recorded origin: Málaga, Spain
Domain history record: 2010–Present (acquired Magnific AI in 2024)
Background & reputation assessment: Operated by Málaga-based Freepik Company S.L. (founded 2010; acquired Magnific AI in 2024), one of the world’s largest design asset and visual AI platforms.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Seamless cross-domain Google OAuth handoff that preserves and resumes the pre-login prompt, rendering 4 parallel candidates via ak-data.magnific.com/app/api/render/v4.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Individual web subscriptions at the current Magnific destination; excludes Teams and separate API · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Premium2026-10-20 ↗

USD 20/month

USD 174/year

Monthly 20,000 general credits/month

Annual 240,000 general credits/year; valid for a year without monthly resets Rules [1]

Premium+2026-10-20 ↗

USD 45/month

USD 405/year

Monthly 45,000 general credits/month

Annual 600,000 general credits/year; valid for a year without monthly resets Rules [1]

Pro Starter2026-10-20 ↗

USD 110/month

USD 990/year

Monthly 112,500 general credits/month

Annual 1,500,000 general credits/year; valid for a year without monthly resets Rules [1]

Pro high volume2026-10-20 ↗

USD 280/month

USD 2,520/year

Monthly 300,000 general credits/month

Monthly price before tax; credits do not establish removal counts

Annual 4,000,000 general credits/year; valid for one year, no monthly resets Rules [1]

8 plan / billing-period records, all listed; — means that period is not recorded.

[1] Annual price before VAT/local taxes, which depend on billing region. Renews annually; renewal can be canceled in subscription settings.

Both individual Pro volumes and all annual charges are verified. Auto-refill was not enabled; the public pricing page does not quote top-up unit prices, free removal limits or per-removal credit cost, so those remain unconfirmed. Prices exclude VAT/local taxes.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Cancel auto-renewal anytime in account subscription settings

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗

Official subscriptions and annual checkout ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗ · Capture 4 ↗ · Capture 5 ↗ · Capture 6 ↗

Allowance & disclosure on test date

Requires free login to use daily free allowance; default Auto model generates 4 candidates per batch and exports unwatermarked PNGs

Prompts for login after clicking generate; workspace defaults to 16:9 and AI prompt rewriting unless manually switched to 1:1 and OFF

Evaluation basis & full scores · Overall 85.9/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 81/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	26 / 35	Clicking generate on the Freepik AI Image Generator landing page redirects to the Magnific workspace (magnific.com/app/ai-image-generator) and prompts unauthenticated visitors to sign in; after free Google sign-in, the pending generation resumes automatically using the daily free allowance to produce a 4-image batch and export an unwatermarked 1024×1024 PNG (though login requirement and exact daily credit rules are not disclosed upfront on the landing hero).	Evidence B →
Interaction restraint	23 / 30	Free download delivers a clean 1024×1024 native PNG without watermarks or credit-card walls; however, the workspace defaults to 16:9 aspect ratio with "AI prompt" rewriting enabled, requiring manual adjustment to 1:1 and turning "AI prompt" OFF for faithful square prompt adherence.	Evidence A →
Background & reputation	19 / 20	Operated by Málaga-based Freepik Company S.L. (founded 2010; acquired Magnific AI in 2024), one of the world’s largest design asset and visual AI platforms.	Evidence B →
Engineering & privacy	13 / 15	Seamless cross-domain Google OAuth handoff that preserves and resumes the pre-login prompt, rendering 4 parallel candidates via ak-data.magnific.com/app/api/render/v4.	Evidence A →

Capability · 88/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	56 / 65	Text Rendering (25.0/25: Exact two-line white "REALTEST 2026" text on the wooden sign with 100% spelling accuracy (25/25)) · Count & Color Binding (15.0/20: 2 blue ceramic mugs on the right are accurate without color bleeding, but only 2 red apples are generated on the left instead of the requested 3 (15/20)) · 5-Finger Hand & Structure (16.0/20: Vintage brass robotic hand holding the sign shows 4 long articulated fingers in front with the thumb obscured behind the board; solid 1:1 composition and warm lighting (16/20))	Evidence A →
Deliverable specifications	20 / 20	Clean unwatermarked export (8/8); exports the first candidate from the standard 1:1 batch as a native 1024×1024 RGB PNG (1,585,669 bytes, watermark-free) (12/12)	Evidence A →
Processing speed	12 / 15	2026-10-21 single-sample network timing under standard conditions (Auto model, explicit 1:1, AI prompt OFF): pure server inference from POST /app/api/start-tti-v2 across 4 parallel POST /app/api/render/v4 calls to POST /app/api/v2/ai/simulate-generation completes in 5.76s (end-to-end 6.12s) (12.0/15)	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

PicWish

picwish.com · Image generation

85/100

Seven-aspect task score

Free login grants 3 daily free generations (0/image) with native 2048×2048 watermark-free JPEG export and accurate sign text and object counts, though it redirects to a chat workspace and defaults to 3 concurrent models that use all 3 daily slots at once.

Tested export JPEG · Output attached

Historical test · 2026-10-22
Signed-in flow; paid tier not standardized · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (JPEG · Output attached) ↗ Entry & rules ↗Result screen ↗Download & limits ↗Additional screenshot 1 ↗

Use this capability ↗Tool profile →Full report →

Task quality

"REALTEST 2026" carved lettering plus 3 red apples and 2 blue ceramic mugs are 100% accurate; the robot thumb sits behind the sign leaving 4 fingers visible in front.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

"REALTEST 2026" carved lettering plus 3 red apples and 2 blue ceramic mugs are 100% accurate; the robot thumb sits behind the sign leaving 4 fingers visible in front.

Original historical observation

With explicit 1:1 aspect ratio and default 2K resolution, PicWish AI Design Create generated a 2048×2048 watermark-free JPEG in 28.99s: "REALTEST 2026" sign text is 100% accurate (25/25) and the 3 red apples + 2 blue ceramic mugs match count and color constraints (20/20), while the robot thumb sits behind the sign leaving 4 fingers visible in front (61.0/65).

Suitable uses · historical assessment

Creators who want to generate 2048×2048 (2K) watermark-free images using 3 free daily generations with accurate English sign text and multi-object count binding

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

JPEG · Output attached · No watermark in test

Detailed facts & historical record

Recorded export specifications: JPEG · Output attached · No watermark in test
Historical specification note · not an official maximum: 2048×2048 (Native 2K watermark-free JPEG)
Watermark observation: Completely watermark-free (2048×2048 clean JPEG export)
Batch & API: Supports switching or concurrent comparison across Nano Banana, GPT Image-2.5, and Seedream models, with one-click handoff to background removal, enhancement, retouching, and outpainting
Recorded free-tier dimensions: 2048 × 2048
Attached file dimensions: Not verified

Evidence & scope →

Speed & reliability

About 29.25s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 28.99s · End-to-end reference: 29.25s
Date & fixture: 2026-10-22 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Free sign-in via Google / Apple / Facebook / Email (redirected from ai-art-generator to ai-design-create chat workspace)
Manual adjustment: Not verified
Interaction observations: Not verified

Evidence & scope →

Background & reputation

Wangxu Technology (HK) Co., Ltd. (Apowersoft / PicWish) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Wangxu Technology (HK) Co., Ltd. (Apowersoft / PicWish)
Recorded origin: Hong Kong / Shenzhen, China
Domain history record: 2021–Present
Background & reputation assessment: Not verified
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Not verified
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Web free, weekly/monthly/annual subscriptions and expanded packs; excludes API · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free2026-10-20 ↗

Free

Free Limited non-HD downloads; no included credits or batch processing

Pro2026-10-20 ↗

USD 5.99/week

Other period

Other period 50 credits/week; 50 core-feature HD downloads/day Rules [1]

Pro2026-10-20 ↗

USD 9.99/month

USD 59.99/year

Monthly 350 credits/month; 50 core-feature HD downloads/day Rules [1]

Annual 350 credits/month; 50 core-feature HD downloads/day Rules [1]

Max2026-10-20 ↗

USD 26.99/month

USD 263.99/year

Monthly 2,000 credits/month; 15,000 HD downloads/month Rules [1]

Annual 2,000 credits/month; 15,000 HD downloads/month Rules [1]

300 credits2026-10-20 ↗

USD 12.99 · one-time payment

Usage / one-time

Usage / one-time 300 credits; 1 per removal; not for API Rules [2]

1,000 credits2026-10-20 ↗

USD 33.99 · one-time payment

Usage / one-time

Usage / one-time 1,000 credits; 1 per removal; not for API Rules [2]

5,000 credits2026-10-20 ↗

USD 65.99 · one-time payment

Usage / one-time

Usage / one-time 5,000 credits; 1 per removal; not for API Rules [2]

11,500 credits2026-10-20 ↗

USD 99.95 · one-time payment

Usage / one-time

Usage / one-time 11,500 credits; 1 per removal; not for API Rules [2]

30,000 credits2026-10-20 ↗

USD 269.99 · one-time payment

Usage / one-time

Usage / one-time 30,000 credits; 1 per removal; not for API Rules [2]

80,000 credits2026-10-20 ↗

USD 639.99 · one-time payment

Usage / one-time

Usage / one-time 80,000 credits; 1 per removal; not for API Rules [2]

Custom credits2026-10-20 ↗

Contact sales

Quoted

Quoted Custom high-volume quote

13 plan / billing-period records, all listed; — means that period is not recorded.

[1] Subscription credits do not roll over; downloads and credits are separate allowances

[2] No expiry

Actual annual charges $59.99/$263.99. A 3-day trial is offered; cancel before expiry to avoid charges. Exact daily free count is unspecified. The 11,500-credit pack includes a current 1,500-credit bonus.

Cancellation, sources & verification record

Cancel under My Orders; request refunds through official support, eligibility not separately verified.

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗ · Capture 4 ↗

Allowance & disclosure on test date

Signed-in free accounts receive 3 free images per day ("You have 3 free images left today" / "0/image"); note that the default "3 models" setting consumes all 3 daily free images concurrently in one prompt unless switched to a single model

Landing hero does not pre-disclose the workspace login redirect; once signed in, the prompt bar clearly displays the remaining 3 daily free images and 0/image cost

Evaluation basis & full scores · Overall 84.9/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 80/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	25 / 35	Explanation unavailable	Evidence A →
Interaction restraint	25 / 30	Explanation unavailable	Evidence A →
Background & reputation	17 / 20	Explanation unavailable	Evidence A →
Engineering & privacy	13 / 15	Explanation unavailable	Evidence A →

Capability · 87/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	61 / 65	Text Rendering (25/25: Two-line carved dark text "REALTEST 2026" on the wooden sign is 100% accurately spelled with crisp woodgrain engraving (25/25)) · Count & Color Binding (20/20: Exactly 3 red apples on the left and 2 blue glazed ceramic mugs on the right on the dark oak table with zero color bleeding (20/20)) · 5-Finger Hand & Structure (16/20: The vintage brass robot holds the wooden sign with 4 clearly articulated mechanical fingers visible on the front (thumb behind the sign), with strong 1:1 studio composition and warm lighting (16/20))	Evidence A →
Deliverable specifications	20 / 20	Forced watermark in free export (0/8); Export & Max deliverable spec: 2048×2048 (Native 2K watermark-free JPEG)	Evidence A →
Processing speed	6 / 15	Network-observed server wait: 28.99s (API mode: async_verified; end-to-end 29.25s shown for reference only, never penalized)	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Fotor

fotor.com · Image generation

83/100

Seven-aspect task score

The first free watermark-free image downloads without login in a simple flow; the second attempt needs signup, and this result split the text and miscounted objects.

Tested export 880×1168 (~1.03MP unwatermarked PNG, 1.13MB) · Historical description; output not attached

Historical test · 2026-10-16
Guest flow · Fixed English prompt · Historical page record; output not attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Spelling was correct, but text placement differed and one extra apple appeared; inspect complex constraints.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Spelling was correct, but text placement differed and one extra apple appeared; inspect complex constraints.

Original historical observation

Strong reverse bias compared to its middling Background Remover (65.1): Fotor is the only major suite tested that delivers an unwatermarked ~1MP guest T2I export (Kling O1) without requiring login.

Suitable uses · historical assessment

Users wanting an instant, zero-login, unwatermarked ~1MP AI image generation (Kling O1) without registering an account

Result & output evidence

No native output obtained; the preview does not establish final export quality.

Open result screenshot ↗

Full test record →

Evidence & scope →

Delivery capabilities

880×1168 (~1.03MP unwatermarked PNG, 1.13MB) · Historical description; output not attached · Check watermark record

Detailed facts & historical record

Recorded export specifications

880×1168 (~1.03MP unwatermarked PNG, 1.13MB) · Historical description; output not attached · Check watermark record

No native output is attached; check the evidence for specification and watermark claims.

Historical specification note · not an official maximum

880×1168 (~1.03MP Unwatermarked PNG · Kling O1)

Watermark observation

Zero watermark (clean PNG export)

Batch & API

Supports multiple models (Kling O1, Flux, Fotor M2) and aspect ratios in workspace

Recorded free-tier dimensions

Not verified

Attached file dimensions

Not verified

Evidence & scope →

Speed & reliability

About 16.5s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 14.8s · End-to-end reference: 16.5s
Date & fixture: 2026-10-16 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

No login in this run

Detailed facts & historical record

Account requirements: No login in this test
Sign-in methods: No login required for the 1st generation & download (login required from 2nd generation onward)
Manual adjustment: Not verified
Interaction observations: Delivers a genuine no-login first generation and direct PNG download, though the page requests browser push notification permissions on load and ignores the "1:1 square" prompt text unless the aspect ratio selector is manually changed from its default 3:4 portrait setting.

Evidence & scope →

Background & reputation

Everimaging Ltd. (Fotor) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Everimaging Ltd. (Fotor)
Recorded origin: Chengdu, China / Global Operations
Domain history record: 2012–present
Background & reputation assessment: Operated by Everimaging Ltd. (Fotor, founded 2012), one of the largest cross-platform online photo suites.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Routes guest generation to the Kling O1 model backend, returning an 880×1168 (1.13MB) unwatermarked PNG in 14.8s.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Web personal free, all Pro/Pro+/Max monthly/annual tiers, credit subscriptions and 10 retail packs; excludes headshots/API · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Basic2026-10-20 ↗

Free

Free Official comparison lists Free forever; removal limits need separate verification

Pro2026-10-20 ↗

USD 10.99/month

USD 47.99/year

Annual 100 general credits/month; batch background-removal tool included

Monthly 100 general credits/month

Pro+2026-10-20 ↗

USD 23.99/month

USD 69.99/year

Annual 300 general credits/month; batch background-removal tool included

Monthly 300 general credits/month

Max 1k2026-10-20 ↗

USD 49.99/month

USD 167.92/year

Annual 1,000 general credits/month; batch background-removal tool included

Monthly 1,000 general credits/month

Credit subscription 2002026-10-20 ↗

USD 13.99/month

USD 69.99/year

Annual 200 credits/month, billed annually Rules [1]

Monthly 200 general credits/month Rules [3]

Credit subscription 5002026-10-20 ↗

USD 25.99/month

USD 143.99/year

Annual 500 credits/month, billed annually Rules [1]

Monthly 500 general credits/month Rules [3]

Credit subscription 10002026-10-20 ↗

USD 49.99/month

USD 259.99/year

Annual 1000 credits/month, billed annually Rules [1]

Monthly 1,000 general credits/month Rules [3]

Credit subscription 20002026-10-20 ↗

USD 99.99/month

USD 459.99/year

Annual 2000 credits/month, billed annually Rules [1]

Monthly 2,000 general credits/month Rules [3]

Credit subscription 50002026-10-20 ↗

USD 239.99/month

USD 1,099.99/year

Annual 5000 credits/month, billed annually Rules [1]

Monthly 5,000 general credits/month Rules [3]

Credit subscription 100002026-10-20 ↗

USD 459.99/month

USD 1,999.99/year

Annual 10000 credits/month, billed annually Rules [1]

Monthly 10,000 general credits/month Rules [3]

Max 2k2026-10-20 ↗

USD 99.99/month

USD 335.92/year

Monthly 2,000 general credits/month

Annual 2,000 general credits/month Rules [2]

Max 5k2026-10-20 ↗

USD 129.99/month

USD 755.92/year

Monthly 5,000 general credits/month

Annual 5,000 general credits/month Rules [2]

Max 10k2026-10-20 ↗

USD 229.99/month

USD 1,427.92/year

Monthly 10,000 general credits/month

Annual 10,000 general credits/month Rules [2]

Max 20k2026-10-20 ↗

USD 479.99/month

USD 2,687.92/year

Monthly 20,000 general credits/month

Annual 20,000 general credits/month Rules [2]

Max 50k2026-10-20 ↗

USD 1,149.99/month

USD 6,719.92/year

Monthly 50,000 general credits/month

Annual 50,000 general credits/month Rules [2]

Retail 1002026-10-20 ↗

USD 19.99 · one-time payment

Usage / one-time

Usage / one-time 100 general credits Rules [4]

Retail 2002026-10-20 ↗

USD 23.99 · one-time payment

Usage / one-time

Usage / one-time 200 general credits Rules [4]

Retail 5002026-10-20 ↗

USD 49.99 · one-time payment

Usage / one-time

Usage / one-time 500 general credits Rules [4]

Retail 10002026-10-20 ↗

USD 79.99 · one-time payment

Usage / one-time

Usage / one-time 1,000 general credits Rules [4]

Retail 20002026-10-20 ↗

USD 119.99 · one-time payment

Usage / one-time

Usage / one-time 2,000 general credits Rules [4]

Retail 50002026-10-20 ↗

USD 259.99 · one-time payment

Usage / one-time

Usage / one-time 5,000 general credits Rules [4]

Retail 100002026-10-20 ↗

USD 499.99 · one-time payment

Usage / one-time

Usage / one-time 10,000 general credits Rules [4]

Retail 200002026-10-20 ↗

USD 899.99 · one-time payment

Usage / one-time

Usage / one-time 20,000 general credits Rules [4]

Retail 300002026-10-20 ↗

USD 1,299.99 · one-time payment

Usage / one-time

Usage / one-time 30,000 general credits Rules [4]

Retail 500002026-10-20 ↗

USD 1,999.99 · one-time payment

Usage / one-time

Usage / one-time 50,000 general credits Rules [4]

39 plan / billing-period records, all listed; — means that period is not recorded.

[1] Rollover while subscribed, capped at 5 months

[2] Observed promotional annual total; subsequent renewal amount not specified on the price card

[3] Rollover while subscribed up to 5 months

[4] Only available to credit subscribers or Pro/Pro+/Max members; retail expiry is not stated in this section

All six Max tiers, monthly credit subscriptions and ten retail packs verified; retail packs require paid membership. Annual amounts preserve the observed promotion. Price cards do not specify renewal quotes, free removal counts, removal conversion or retail expiry; these remain unconfirmed.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Cancel auto-renewal anytime in account settings

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗

All subscription tiers and retail credits ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗ · Capture 4 ↗ · Capture 5 ↗ · Capture 6 ↗ · Capture 7 ↗ · Capture 8 ↗ · Capture 9 ↗ · Capture 10 ↗ · Capture 11 ↗ · Capture 12 ↗ · Capture 13 ↗ · Capture 14 ↗

Allowance & disclosure on test date

1 free guest generation without signup via Kling O1 model; subsequent generations require login and credits

No-signup trial works immediately; login prompt appears only when attempting a second generation

Evaluation basis & full scores · Overall 83.3/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 84/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	28 / 35	Page title promises "Free AI Image Generator: Text to Image Online (No Signup)"; real-machine testing confirms guests can generate and download their 1st image (880×1168 unwatermarked PNG via Kling O1) with zero login, while a login modal appears on the 2nd attempt.	Evidence A →
Interaction restraint	25 / 30	Delivers a genuine no-login first generation and direct PNG download, though the page requests browser push notification permissions on load and ignores the "1:1 square" prompt text unless the aspect ratio selector is manually changed from its default 3:4 portrait setting.	Evidence A →
Background & reputation	18 / 20	Operated by Everimaging Ltd. (Fotor, founded 2012), one of the largest cross-platform online photo suites.	Evidence C →
Engineering & privacy	13 / 15	Routes guest generation to the Kling O1 model backend, returning an 880×1168 (1.13MB) unwatermarked PNG in 14.8s.	Evidence A →

Capability · 83/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	54 / 65	Text Rendering (22.5/25: Creative split rendering: "REALTEST" is carved accurately on the wooden sign while "2026" is printed on a giant 5-fingered hand mounted on the robot's left arm (22.5/25)) · Count & Color Binding (15/20: 2 blue ceramic mugs are exact, but generated 4 red apples instead of 3 (+1 extra apple) (15.0/20)) · 5-Finger Hand & Structure (16.5/20: Brass right hand holding the sign shows 4 mechanical fingers, while the secondary left hand displays 5 distinct fingers stamped with "2026" (16.5/20))	Evidence B →
Deliverable specifications	19 / 20	Clean unwatermarked export (8/8); Export & Max deliverable spec: 880×1168 (~1.03MP Unwatermarked PNG · Kling O1)	Evidence B →
Processing speed	10 / 15	Network-observed server wait: 14.8s (API mode: async_poll; end-to-end 16.5s shown for reference only, never penalized)	Evidence B →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

LightX

lightxeditor.com · Image generation

83/100

Seven-aspect task score

Requires sign-in and manually switching from the default 8-credit model to 2-credit GPT Image 1.5 Low to use daily free credits; generates in ~5.6s with accurate text and object counts, but stamps a bottom-left watermark on the 1024×1024 JPEG.

Tested export JPEG · Output attached

Historical test · 2026-10-23
Signed-in flow; paid tier not standardized · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (JPEG · Output attached) ↗ Entry & rules ↗Result screen ↗Download & limits ↗Additional screenshot 1 ↗

Use this capability ↗Tool profile →Full report →

Task quality

"REALTEST 2026" sign text plus 3 red apples and 2 blue ceramic mugs are 100% accurate; the robot grips the sign with 4 visible fingers per hand and a semi-transparent LightX watermark at the bottom-left.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Original historical observation

Switched to GPT Image 1.5 Low (2 credits), LightX generated a 1024×1024 JPEG in 5.57s: "REALTEST 2026" sign text is 100% accurate (25/25), 3 red apples and 2 blue ceramic mugs match count and color constraints (20/20), and the robot grips the sign with 4 visible fingers per hand (16/20), with a bottom-left LightX watermark on the free export.

Suitable uses · historical assessment

Users seeking fast generation (~5.6s), accurate English text rendering, and exact object count binding who do not mind a small bottom-left watermark on the free tier

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

JPEG · Output attached · Check watermark record

Detailed facts & historical record

Recorded export specifications: JPEG · Output attached · Check watermark record
Historical specification note · not an official maximum: 1024×1024 (JPEG, 266.2 KB)
Watermark observation: Semi-transparent "LightX" brand watermark in the bottom-left corner
Batch & API: Supports multiple model options (GPT Image 1.5 Low, Flux, LightX models) and commercial API
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: Not verified

Evidence & scope →

Speed & reliability

About 6.4s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 5.57s · End-to-end reference: 6.4s
Date & fixture: 2026-10-23 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Supports free Google / Facebook / Apple / Email sign-in (6 daily free AI credits granted)
Manual adjustment: Not verified
Interaction observations: The workspace clearly labels credit costs per model (e.g., 2 credits for GPT Image 1.5 Low) and aspect ratio dimensions (Square 1:1 1024x1024 px, 02_workspace.png). Once generated, clicking "More options -> Download" (03_download.png) on the result card directly downloads the 1024×1024 JPEG without card-binding popups.

Evidence & scope →

Background & reputation

AndOr Communications Pvt. Ltd. (LightX) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: AndOr Communications Pvt. Ltd. (LightX)
Recorded origin: New Delhi, India
Domain history record: 2017–Present
Background & reputation assessment: Operated by AndOr Communications Pvt. Ltd. (New Delhi, India), creator of the LightX mobile and web visual AI suite with tens of millions of users.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Invokes cloud text-to-image inference via POST https://www.instagraphe.mobi/andor-media-1.0/aiartweb/generateImage; following vendor backend storage credential recovery on 2026-10-23, single-sample server wait measured 5.57s (6.40s end-to-end).
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

USD web monthly/annual subscriptions and 800-credit pack; excludes API · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Starter2026-10-20 ↗

Free

Free Card states 6 AI credits/day; comparison table instead says up to 5 credits

Background removal outputs low-resolution JPG; export table lists 512px

Pro2026-10-20 ↗

USD 9.99/month

USD 59.99/year

Annual 1,000 AI credits/month; unlimited background removal/replacement

PNG and high resolution; export table lists 4096px

Monthly 1,000 AI credits/month; unlimited background removal/replacement

PNG and high resolution; export table lists 4096px

800 AI credit pack2026-10-20 ↗

USD 9.99 · one-time payment

Usage / one-time

Usage / one-time 800 general AI credits

Advertised lifetime validity

4 plan / billing-period records, all listed; — means that period is not recorded.

Monthly $9.99 verified; promotional annual total $59.99 and non-expiring 800-credit pack $9.99. The free card still says 6 daily credits while the table says up to 5. Free removal is low-resolution JPG; Pro provides unlimited high-resolution PNG removal. Earlier INR and current USD observations do not imply identical regional pricing.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Cancellation and refund terms follow official checkout and help policies.

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗

Monthly pricing and feature comparison ↗ · 2026-10-20 Capture 1 ↗

Allowance & disclosure on test date

Free accounts receive 6 daily credits, which cannot run the default 8-credit model; users must switch to GPT Image 1.5 Low (2 credits/generation) to generate and download a 1024×1024 JPEG

Model selector clearly shows credit cost per generation, but does not pre-disclose that free exports carry a bottom-left watermark

Evaluation basis & full scores · Overall 83.2/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 73/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	20 / 35	The landing page (01_landing.png) requires sign-in before entering the text-to-image workspace; free accounts receive 6 daily credits, which is insufficient for the default 8-credit model. Switching manually to "GPT Image 1.5 Low" (labeled 1 generation = 2 credits, 02_workspace.png) allows free generation, though the exported 1024×1024 JPEG includes a bottom-left "LightX" brand watermark.	Evidence A →
Interaction restraint	23 / 30	The workspace clearly labels credit costs per model (e.g., 2 credits for GPT Image 1.5 Low) and aspect ratio dimensions (Square 1:1 1024x1024 px, 02_workspace.png). Once generated, clicking "More options -> Download" (03_download.png) on the result card directly downloads the 1024×1024 JPEG without card-binding popups.	Evidence A →
Background & reputation	18 / 20	Operated by AndOr Communications Pvt. Ltd. (New Delhi, India), creator of the LightX mobile and web visual AI suite with tens of millions of users.	Evidence B →
Engineering & privacy	12 / 15	Invokes cloud text-to-image inference via POST https://www.instagraphe.mobi/andor-media-1.0/aiartweb/generateImage; following vendor backend storage credential recovery on 2026-10-23, single-sample server wait measured 5.57s (6.40s end-to-end).	Evidence A →

Capability · 87.5/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	61 / 65	Text rendering (25.0/25: two-line carved lettering "REALTEST 2026" on the wooden sign is 100% accurately spelled and crisp) · Count & color binding (20.0/20: 3 red apples on the left and 2 blue ceramic mugs on the right match exact count and color constraints) · 5-finger anatomy & composition (16.0/20: vintage brass robot grips the upper corners of the sign with 4 visible mechanical fingers per hand, warm studio composition) (61.0/65)	Evidence A →
Deliverable specifications	14 / 20	Free tier (GPT Image 1.5 Low, 2 credits) exports a native 1024×1024 RGB JPEG (266,153 bytes, 12/12 resolution score), but stamps a semi-transparent "LightX" watermark at the bottom-left (2/8 watermark score) (14.0/20)	Evidence A →
Processing speed	12.5 / 15	2026-10-23 same-run single-sample network timing: POST https://www.instagraphe.mobi/andor-media-1.0/aiartweb/generateImage (r2) server wait measured 5.57s (6.40s end-to-end including CDN image load) (12.5/15)	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

inPixio

inpixio.com · Image generation

82/100

Seven-aspect task score

New free accounts receive 50 AI credits; with default Prompt enhance toggled off, a 4-image batch costs 20 credits, returns in ~9.9s, and exports a 1024×1024 watermark-free PNG with exact sign text but 4 red apples instead of 3.

Tested export 1024 × 1024 · PNG

Historical test · 2026-10-23
Signed-in flow; paid tier not standardized · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (1024 × 1024 · PNG) ↗ Entry & rules ↗Result screen ↗Download & limits ↗Additional screenshot 1 ↗

Use this capability ↗Tool profile →Full report →

Task quality

With Prompt enhance off, "REALTEST 2026" sign lettering and 2 blue ceramic mugs are exact, whereas the plate holds 4 red apples instead of 3 and 4 metallic fingers are visible gripping the sign.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

With Prompt enhance off, "REALTEST 2026" sign lettering and 2 blue ceramic mugs are exact, whereas the plate holds 4 red apples instead of 3 and 4 metallic fingers are visible gripping the sign.

Original historical observation

After disabling the default Prompt enhance toggle and submitting the unmodified canonical prompt, 4 concurrent requests returned four 1024×1024 candidates in 9.89s: Candidate 1 renders "REALTEST 2026" with 100% accuracy (25/25) and 2 blue ceramic mugs alongside 4 red apples (16/20), exporting as a watermark-free PNG.

Suitable uses · historical assessment

Users who want to use 50 free new-account credits to generate multiple candidate images at once and export 1024×1024 watermark-free PNGs

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

1024 × 1024 · PNG · No watermark in test

Detailed facts & historical record

Recorded export specifications: 1024 × 1024 · PNG · No watermark in test
Historical specification note · not an official maximum: 1024×1024 (watermark-free native PNG, 1.62 MB)
Watermark observation: Completely watermark-free (native 1024×1024 PNG export)
Batch & API: Supports 1–4 images per batch, preset style selectors, Prompt enhance toggle, and one-click handoff to the online photo editor
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: 1024 × 1024 · PNG

Evidence & scope →

Speed & reliability

About 11.13s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 9.89s · End-to-end reference: 11.13s
Date & fixture: 2026-10-23 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Supports free Google or email sign-in (new free accounts receive 50 free AI credits and 5 free downloads)
Manual adjustment: Not verified
Interaction observations: The workspace enables "Prompt enhance" by default (which rewrites prompts) and defaults to 4 images per batch (20 credits), requiring users to toggle Prompt enhance off or adjust the image count manually (02_workspace_canonical.png); once signed in, generating and downloading a 1024×1024 watermark-free PNG proceeds without paywall popups (04_download_submenu.png).

Evidence & scope →

Background & reputation

Avanquest Software SAS (Claranova SE, Paris, France) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Avanquest Software SAS (Claranova SE, Paris, France)
Recorded origin: Paris, France (Avanquest software group)
Domain history record: 10+ year brand history
Background & reputation assessment: Operated by Avanquest Software SAS, a subsidiary of Euronext Paris-listed Claranova SE and a long-standing consumer photo software brand.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Issues 4 concurrent POST https://v2-api.inpixio.com/v1/text-to-image requests returning 1024×1024 images (2026-10-23 single-sample server wait measured 9.89s, 11.13s end-to-end); workspace session loads third-party marketing analytics.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Current official purchase-page free, monthly, annual and all five credit packs · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free2026-10-20 ↗

Free

Free Current purchase page: 50 AI credits/month, 5 HD downloads/month, no watermark

Feature FAQ still says free users cannot save or download HD results, conflicting with the purchase page

Pro2026-10-20 ↗

USD 8.95/month

USD 69.95/year

Monthly 100 AI credits/month; 100 HD downloads/month

No watermark; web and desktop; all features

Annual 100 AI credits/month; 100 HD downloads/month

No watermark; web and desktop; all features

Ultimate2026-10-20 ↗

USD 12.95/month

USD 89.95/year

Monthly 250 AI credits/month; unlimited HD downloads

No watermark; web and desktop; all features

Annual 250 AI credits/month; unlimited HD downloads

No watermark; web and desktop; all features

200 AI credits2026-10-20 ↗

USD 9.95 · one-time payment

Usage / one-time

Usage / one-time 200 general AI credits; applicability to removal is unspecified Rules [1]

500 AI credits2026-10-20 ↗

USD 22.95 · one-time payment

Usage / one-time

Usage / one-time 500 general AI credits; applicability to removal is unspecified Rules [1]

1000 AI credits2026-10-20 ↗

USD 43.95 · one-time payment

Usage / one-time

Usage / one-time 1000 general AI credits; applicability to removal is unspecified Rules [1]

2000 AI credits2026-10-20 ↗

USD 79.95 · one-time payment

Usage / one-time

Usage / one-time 2000 general AI credits; applicability to removal is unspecified Rules [1]

5000 AI credits2026-10-20 ↗

USD 149.95 · one-time payment

Usage / one-time

Usage / one-time 5000 general AI credits; applicability to removal is unspecified Rules [1]

10 plan / billing-period records, all listed; — means that period is not recorded.

[1] No expiry

The recheck preserves a conflict: official purchase cards list 5 free HD downloads/month without a watermark, while the feature FAQ says free users cannot save or download HD. Free HD availability is therefore not guaranteed. AI credits cover AI backgrounds/generation, not established removal counts. Purchase cards state a 30-day money-back guarantee, subject to refund eligibility terms.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

The page allows cancellation/plan changes anytime and advertises a 30-day money-back guarantee, subject to terms.

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗ · Capture 4 ↗

Feature FAQ statement about free HD ↗ · 2026-10-20 Capture 1 ↗

Plan comparison reached through official purchase entry ↗ · 2026-10-20 Capture 1 ↗

Allowance & disclosure on test date

Free registration grants 50 AI credits; text-to-image costs 5 credits per image (default 4-image batch costs 20 credits, adjustable from 1 to 4 images) and supports direct 1024×1024 watermark-free PNG or JPG download

Redirects to sign-in upon clicking generate on the landing page; workspace top bar displays remaining AI credits in real time (e.g., 50 -> 30 -> 10)

Evaluation basis & full scores · Overall 82.1/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 73/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	23 / 35	The landing page (01_landing.png) provides a prompt box and "Generate image" button that redirects to online.inpixio.com/en and requires signing into a free account. Newly registered free accounts clearly show 50 AI credits and 5 free downloads in the top bar; generating the default 4-image batch consumes 20 credits (5 credits/image, 02_workspace_canonical.png, 03_result_canonical.png) and exports a native 1024×1024 watermark-free PNG via "Download -> PNG Suggested" (04_download_submenu.png).	Evidence A →
Interaction restraint	22 / 30	The workspace enables "Prompt enhance" by default (which rewrites prompts) and defaults to 4 images per batch (20 credits), requiring users to toggle Prompt enhance off or adjust the image count manually (02_workspace_canonical.png); once signed in, generating and downloading a 1024×1024 watermark-free PNG proceeds without paywall popups (04_download_submenu.png).	Evidence A →
Background & reputation	17 / 20	Operated by Avanquest Software SAS, a subsidiary of Euronext Paris-listed Claranova SE and a long-standing consumer photo software brand.	Evidence B →
Engineering & privacy	11 / 15	Issues 4 concurrent POST https://v2-api.inpixio.com/v1/text-to-image requests returning 1024×1024 images (2026-10-23 single-sample server wait measured 9.89s, 11.13s end-to-end); workspace session loads third-party marketing analytics.	Evidence A →

Capability · 86/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	56 / 65	Text rendering (25.0/25: with Prompt enhance disabled, single-line stencil lettering "REALTEST 2026" on the wooden sign is 100% accurately spelled and sharp) · Count & color binding (16.0/20: 2 blue ceramic mugs on the right are exact, while the plate on the left contains 4 red apples instead of 3, with zero color bleeding) · 5-finger anatomy & composition (15.0/20: vintage brass robot grips the left edge of the sign with 4 visible metallic fingers, natural warm studio composition) (56.0/65)	Evidence A →
Deliverable specifications	20 / 20	Watermark-free export (8/8); under free account credits, Download -> PNG (Suggested) directly exports a native 1024×1024 RGB PNG (1,621,663 bytes, zero watermark) (12/12) (20.0/20)	Evidence A →
Processing speed	10 / 15	2026-10-23 same-run single-sample network timing: 4 concurrent POST https://v2-api.inpixio.com/v1/text-to-image requests (r6..r9, fromServiceWorker: false) measured 9.89s server wait (0.60s response transfer, 11.13s end-to-end) (10.0/15)	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

MagicEraser.org

magiceraser.org · Image generation

82/100

Seven-aspect task score

Login-free trials have clear credit costs and support further editing on the same page; credits are limited, and this 1000px result has spelling and count errors.

Tested export JPEG · Output attached

Historical test · 2026-10-16
Guest flow · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (JPEG · Output attached) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Materials looked plausible, but one letter was wrong and one apple missing; check exact text and counts.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Materials looked plausible, but one letter was wrong and one apple missing; check exact text and counts.

Original historical observation

MagicEraser 6.0 renders convincing vintage brass textures and carved wood grain with "REALLEST 2026" (1-letter typo) and 2 blue mugs, though it rendered 2 red apples instead of 3.

Suitable uses · historical assessment

Creators who want no-login text-to-image generation combined with inpainting and upscaling in one workspace

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

JPEG · Output attached · No watermark in test

Detailed facts & historical record

Recorded export specifications: JPEG · Output attached · No watermark in test
Historical specification note · not an official maximum: 1000×1000 (Guest no-login unwatermarked JPG export)
Watermark observation: No watermark
Batch & API: Supports batchMode toggle
Recorded free-tier dimensions: 1000 × 1000
Attached file dimensions: Not verified

Evidence & scope →

Speed & reliability

About 9.36s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 8.96s · End-to-end reference: 9.36s
Date & fixture: 2026-10-16 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

No login in this run

Detailed facts & historical record

Account requirements: No login in this test
Sign-in methods: No login required (guests receive 10 Free Credits automatically)
Manual adjustment: Directly switch to Erase or Fine Edit on the generated image
Interaction observations: One-click Download button exports the unwatermarked JPG and allows instant switching to Erase, Enhance, or Upscaler in the same workspace.

Evidence & scope →

Background & reputation

MagicEraser.org / LinoCut AI · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: MagicEraser.org / LinoCut AI
Recorded origin: US / APAC Independent AI Team
Domain history record: 2023–2026
Background & reputation assessment: Part of the MagicEraser.org / LinoCut AI suite sharing a unified guest credit pool.
External reputation record: One of the very few AI Image Generators that allows guests to generate and download 1000×1000 unwatermarked images without logging in.

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Claims deletion within 4 hours; deletion not independently verified

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Google Analytics · Aliyun SLS Web Tracking
Engineering & privacy assessment: Executes via HTTPS async polling (me_single_photo_editor_v1) in 8.96s and returns a signed US-West OSS image.
Retention statement: Claims deletion within 4 hours; deletion not independently verified

Evidence & scope →

Pricing & rules

Web subscriptions, displayed top-ups and enterprise quote; USD · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free2026-10-20 ↗

Free

Free Free use varies by tool and available allowance

200 Credit2026-10-20 ↗

USD 8.9/month

—

Monthly 200 general credits/month Rules [1] [2]

500 Credit2026-10-20 ↗

USD 12.9/month

—

Monthly 500 general credits/month Rules [1] [2]

1000 Credit2026-10-20 ↗

USD 16.9/month

—

Monthly 1000 general credits/month Rules [1] [2]

200 Credit2026-10-20 ↗

USD 14.99 · one-time payment

Usage / one-time

Usage / one-time 200 general credits Rules [3]

500 Credit2026-10-20 ↗

USD 30 · one-time payment

Usage / one-time

Usage / one-time 500 general credits Rules [3]

2000 Credit2026-10-20 ↗

USD 70 · one-time payment

Usage / one-time

Usage / one-time 2000 general credits Rules [3]

5000 Credit2026-10-20 ↗

USD 90 · one-time payment

Usage / one-time

Usage / one-time 5000 general credits Rules [3]

Enterprise2026-10-20 ↗

Contact sales

Quoted

Quoted Custom allowances, API, team and support

10,000 credit pack2026-10-20 ↗

USD 120 · one-time payment

Usage / one-time

Usage / one-time 10,000 cross-tool credits Rules [4]

30,000 credit pack2026-10-20 ↗

USD 330 · one-time payment

Usage / one-time

Usage / one-time 30,000 cross-tool credits Rules [4]

11 plan / billing-period records, all listed; — means that period is not recorded.

[1] Auto-renews; cancel anytime

[2] Use subscription credits before expiry

[3] No expiry

[4] Never expires

Verified 10,000 credits for $120 and 30,000 for $330, both one-time payments with non-expiring credits. The pricing page quantifies upscaling, restoration, editing and video use but not background removal; general credits are not asserted removal counts.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Cancel anytime in account dashboard

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗

Large top-up package selections ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗

Allowance & disclosure on test date

Guests get 10 Free Credits without login; MagicEraser 6.0 Text-to-Image costs only 1 Credit per generation (up to 10 free unwatermarked images)

Displays "Credits Required: 1 Credit" directly above the Generate button

Evaluation basis & full scores · Overall 81.9/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 92/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	33 / 35	Clearly displays 10 initial guest Free Credits in the top bar and states "Credits Required: 1 Credit" on the Text-to-Image panel; generates and exports a 1000×1000 unwatermarked JPG without login.	Evidence A →
Interaction restraint	28 / 30	One-click Download button exports the unwatermarked JPG and allows instant switching to Erase, Enhance, or Upscaler in the same workspace.	Evidence A →
Background & reputation	18 / 20	Part of the MagicEraser.org / LinoCut AI suite sharing a unified guest credit pool.	Evidence C →
Engineering & privacy	13 / 15	Executes via HTTPS async polling (me_single_photo_editor_v1) in 8.96s and returns a signed US-West OSS image.	Evidence B →

Capability · 77.6/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	48 / 65	Text Rendering (18/25: Two-line sign clearly legible as "REALLEST 2026" (single-letter L/T substitution, 18/25)) · Count & Color Binding (14/20: 2 red apples (1 short of 3) and 2 blue ceramic mugs on the dark oak table with zero color bleeding (14/20)) · 5-Finger Hand & Structure (16/20: Vintage brass robot holding the wooden sign with both hands: right hand shows 5 distinct metallic fingers, left hand shows 4 fingers gripping the side edge (16/20))	Evidence A →
Deliverable specifications	19.6 / 20	Clean unwatermarked export (8/8); Export & Max deliverable spec: 1000×1000 (Guest no-login unwatermarked JPG export)	Evidence A →
Processing speed	10 / 15	Network-observed server wait: 8.96s (API mode: async_polling; end-to-end 9.36s shown for reference only, never penalized)	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Let's Enhance

letsenhance.io · Image generation

80/100

Seven-aspect task score

Upfront credit costs and varied presets accompany good text and counts; signup is required, free 2K images retain a watermark, and this run took about 32 seconds.

Tested export 2048×2048 (9.96 MB PNG · Bottom-left watermark on free credits) · Historical description; output not attached

Historical test · 2026-10-16
Signed-in flow; paid tier not standardized · Fixed English prompt · Historical page record; output not attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Text and counts matched the prompt with 2048px output; the free-credit result carried a watermark.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Text and counts matched the prompt with 2048px output; the free-credit result carried a watermark.

Original historical observation

Top-tier 5-constraint adherence (63.5/65): 100% accurate 'REALTEST 2026' wooden sign text, exact 3 red apples + 2 blue ceramic mugs, and native 2048×2048 (9.96 MB) PNG output, tempered only by a bottom-left watermark on free-credit exports.

Suitable uses · historical assessment

Users who want native 2048×2048 ultra-high-resolution AI image generation coupled with Let's Enhance's super-resolution detail pipeline

Result & output evidence

No native output obtained; the preview does not establish final export quality.

Open result screenshot ↗

Full test record →

Evidence & scope →

Delivery capabilities

2048×2048 (9.96 MB PNG · Bottom-left watermark on free credits) · Historical description; output not attached · Check watermark record

Detailed facts & historical record

Recorded export specifications

2048×2048 (9.96 MB PNG · Bottom-left watermark on free credits) · Historical description; output not attached · Check watermark record

No native output is attached; check the evidence for specification and watermark claims.

Historical specification note · not an official maximum

2048×2048 (PNG · Native Super-Resolution Coupled Output · Bottom-Left Watermark on Free Credits)

Watermark observation

Semi-transparent 'Let's Enhance .io' watermark in the bottom-left corner on free-credit exports

Batch & API

Integrated with one-click 4x–16x upscaling inside `/boost` and Claid.ai enterprise API

Recorded free-tier dimensions

Not verified

Attached file dimensions

Not verified

Evidence & scope →

Speed & reliability

About 32.2s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 31.8s · End-to-end reference: 32.2s
Date & fixture: 2026-10-16 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Mandatory Google or Email sign-in (grants 10 free starter credits)
Manual adjustment: Not verified
Interaction observations: Clean workspace at `/boost` with zero ad popups or countdown traps; clearly displays remaining credit balance ('9 credits left' -> '7 credits left') in the top header.

Evidence & scope →

Background & reputation

Let's Enhance Inc. (San Francisco, CA, USA) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Let's Enhance Inc. (San Francisco, CA, USA)
Recorded origin: San Francisco, USA (Ukrainian-founded team)
Domain history record: 2017–Present
Background & reputation assessment: Operated by Let's Enhance Inc. (San Francisco, CA, founded 2017, Techstars alum), coupling its flagship super-resolution pipeline directly into its text-to-image generator.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Generates and auto-upscales native 2048×2048 (9.96 MB) PNGs in 31.8s server time via Google Cloud Storage (`storage.googleapis.com/production-user-data`).
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Let’s Enhance personal subscription plans (Free Trial / Starter / Pro / Max); USD · 2026-10-21

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Signup free allowance2026-10-21 ↗

Free trial

Trial

Trial 10 free credits upon signup (1 credit per upscaled image, 2 credits per AI generation; free-credit exports include watermark).

Starter Monthly (100 credits)2026-10-21 ↗

USD 12/month

—

Monthly 100 credits/month (unused credits roll over up to 600 credits); upscaling up to 256 MP, AI art generation, and watermark-free export.

Starter Annual (100 credits/mo)2026-10-21 ↗

—

Annual price not verified

Annual 100 credits/month (accumulate up to 600 credits); card displays USD 9/month equivalent billed annually.

Public card displays only the USD 9/month equivalent ("Billed annually") without showing the annual invoice total before login.

Pro Monthly (300 credits)2026-10-21 ↗

USD 32/month

—

Monthly 300 credits/month (unused credits roll over up to 1,800 credits); upscaling up to 350 MP, AI art generation, and watermark-free export.

Pro Annual (300 credits/mo)2026-10-21 ↗

—

Annual price not verified

Annual 300 credits/month (accumulate up to 1,800 credits); card displays USD 24/month equivalent billed annually.

Public card displays only the USD 24/month equivalent ("Billed annually") without showing the annual invoice total before login.

Max Monthly (500 credits)2026-10-21 ↗

USD 45/month

—

Monthly 500 credits/month (unused credits roll over up to 3,000 credits); upscaling up to 512 MP, AI art generation, and watermark-free export.

Max Annual (500 credits/mo)2026-10-21 ↗

—

Annual price not verified

Annual 500 credits/month (accumulate up to 3,000 credits); card displays USD 34/month equivalent billed annually.

Public card displays only the USD 34/month equivalent ("Billed annually") without showing the annual invoice total before login.

7 plan / billing-period records, all listed; — means that period is not recorded.

The official pricing page and FAQ list personal monthly prices: Starter USD 12/month (100 credits/month), Pro USD 32/month (300 credits/month), and Max USD 45/month (500 credits/month), with unused credits rolling over up to 6x the monthly allowance. Annual cards display only monthly equivalents (USD 9 / 24 / 34 per month, "Billed annually") without printing the annual invoice total on the public page, so annual totals are marked unverified.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

Cancel anytime via the Cancel Subscription button on the Pricing page; remaining unused credits expire at the end of the paid billing period. Purchasing any plan or bundle unlocks watermark-free downloads for previously processed images.

Let’s Enhance official pricing and credit rules FAQ ↗ · 2026-10-21 Capture 1 ↗

Allowance & disclosure on test date

10 free credits on signup; AI Image Generation costs 2 credits per 2048×2048 image (up to 5 free 2048px generations)

Credit cost (2 credits) is shown before generation; watermark appears on the generated result and downloaded PNG

Evaluation basis & full scores · Overall 80.2/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 70/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	18 / 35	Landing page `/image-generation` (Screenshot ①) prompts 'Sign up and generate 10 images for free'; signing in with Google grants 10 free credits (each T2I generation costs 2 credits, leaving 7 credits after our 1-credit upscale + 2-credit T2I tests), and outputs a massive 2048×2048 PNG, though free-credit downloads embed a semi-transparent 'Let's Enhance .io' watermark in the bottom-left corner.	Evidence A →
Interaction restraint	21 / 30	Clean workspace at `/boost` with zero ad popups or countdown traps; clearly displays remaining credit balance ('9 credits left' -> '7 credits left') in the top header.	Evidence A →
Background & reputation	18 / 20	Operated by Let's Enhance Inc. (San Francisco, CA, founded 2017, Techstars alum), coupling its flagship super-resolution pipeline directly into its text-to-image generator.	Evidence A →
Engineering & privacy	13 / 15	Generates and auto-upscales native 2048×2048 (9.96 MB) PNGs in 31.8s server time via Google Cloud Storage (`storage.googleapis.com/production-user-data`).	Evidence A →

Capability · 84.5/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	63.5 / 65	Text Rendering (25/25: Exact stencil-carved 'REALTEST 2026' on the wooden sign held by the brass robot, 100% letter-for-letter accurate (25/25)) · Count & Color Binding (20/20: Strictly 3 glossy red apples on the left and 2 blue ceramic mugs on the right on a dark oak table, zero color bleeding (20/20)) · 5-Finger Hand & Structure (18.5/20: Detailed vintage brass robot with rivets and weathered patina gripping the wooden sign with 4 visible front mechanical fingers (18.5/20))	Evidence B →
Deliverable specifications	12 / 20	Clean unwatermarked export (8/8); Export & Max deliverable spec: 2048×2048 (PNG · Native Super-Resolution Coupled Output · Bottom-Left Watermark on Free Credits)	Evidence B →
Processing speed	9 / 15	Network-observed server wait: 31.8s (API mode: async_poll; end-to-end 32.2s shown for reference only, never penalized)	Evidence B →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

media.io

media.io · Image generation

75/100

Seven-aspect task score

Signed-in free Wondershare accounts run Nano Banana 2 Lite (1K, 1:1, 2 credits) in 15.2s server wait with accurate sign lettering and 5-digit hands, but free export is a corner-watermarked JPEG (unwatermarked requires VIP).

Tested export JPEG · Output attached

Historical test · 2026-10-21
Signed-in flow; paid tier not standardized · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (JPEG · Output attached) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

'REALTEST 2026' lettering (25/25) and 5-digit robot hand structure (19/20) are crisp and accurate; 2 blue mugs are exact while 4 red apples were rendered instead of 3 (60.0/65).

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

'REALTEST 2026' lettering (25/25) and 5-digit robot hand structure (19/20) are crisp and accurate; 2 blue mugs are exact while 4 red apples were rendered instead of 3 (60.0/65).

Original historical observation

Used remaining 2 free credits for one Nano Banana 2 Lite image at 1K, 1:1, exact canonical prompt. Generated and downloaded native image with visible Media.io AI Gen watermark. Public visibility default ON shown in screenshot. No purchase. Correction: generation succeeded, but first download click produced no download event. Native export not yet saved; prior observation about downloaded native image was premature. Reopen result to finish download. Completed native download by selecting the Watermark option after opening the download menu. output.jpg is the actual delivered JPEG. Earlier click only opened that menu. Generation used 2 credits and exact 1:1 anchor prompt.

Suitable uses · historical assessment

Users previewing Nano Banana 2 Lite prompt adherence and text rendering who do not mind a bottom-right corner watermark on free exports

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

JPEG · Output attached · Check watermark record

Detailed facts & historical record

Recorded export specifications: JPEG · Output attached · Check watermark record
Historical specification note · not an official maximum: [1024, 1024]
Watermark observation: Visible 'M Media.io AI Gen' watermark in bottom-right corner on free tier (unwatermarked export requires VIP)
Batch & API: Supports multiple models (Nano Banana 2 Lite, Nano Banana Pro, Seedream, etc.), 1–4 outputs per batch, and one-click Agent refinement or video animation
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: Not verified

Evidence & scope →

Speed & reliability

About 15.48s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 15.17s · End-to-end reference: 15.48s
Date & fixture: 2026-10-21 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Detailed facts & historical record

Account requirements: Account required in this test
Sign-in methods: Authorized test profile; see dated run record
Manual adjustment: Not verified
Interaction observations: Historical rubric scores withdrawn pending review of the new run. Used remaining 2 free credits for one Nano Banana 2 Lite image at 1K, 1:1, exact canonical prompt. Generated and downloaded native image with visible Media.io AI Gen watermark. Public visibility default ON shown in screenshot. No purchase. Correction: generation succeeded, but first download click produced no download event. Native export not yet saved; prior observation about downloaded native image was premature. Reopen result to finish download. Completed native download by selecting the Watermark option after opening the download menu. output.jpg is the actual delivered JPEG. Earlier click only opened that menu. Generation used 2 credits and exact 1:1 anchor prompt.

Evidence & scope →

Background & reputation

Publicly Listed Multimedia Software Giant AI Flagship (Wondershare, 300624.SZ) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Publicly Listed Multimedia Software Giant AI Flagship (Wondershare, 300624.SZ)
Recorded origin: Shenzhen, China / Vancouver, Canada
Domain history record: Publicly operating online service
Background & reputation assessment: Historical rubric scores withdrawn pending review of the new run. Used remaining 2 free credits for one Nano Banana 2 Lite image at 1K, 1:1, exact canonical prompt. Generated and downloaded native image with visible Media.io AI Gen watermark. Public visibility default ON shown in screenshot. No purchase. Correction: generation succeeded, but first download click produced no download event. Native export not yet saved; prior observation about downloaded native image was premature. Reopen result to finish download. Completed native download by selecting the Watermark option after opening the download menu. output.jpg is the actual delivered JPEG. Earlier click only opened that menu. Generation used 2 credits and exact 1:1 anchor prompt.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Historical rubric scores withdrawn pending review of the new run. Used remaining 2 free credits for one Nano Banana 2 Lite image at 1K, 1:1, exact canonical prompt. Generated and downloaded native image with visible Media.io AI Gen watermark. Public visibility default ON shown in screenshot. No purchase. Correction: generation succeeded, but first download click produced no download event. Native export not yet saved; prior observation about downloaded native image was premature. Reopen result to finish download. Completed native download by selecting the Watermark option after opening the download menu. output.jpg is the actual delivered JPEG. Earlier click only opened that menu. Generation used 2 credits and exact 1:1 anchor prompt.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Media.io membership plans (Free Trial / Standard / Premium monthly and annual); USD · 2026-10-21

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free Trial2026-10-21 ↗

Free

Free 5 signup credits for new accounts (image upscaling costs 2 credits/run, allowing 2 free unwatermarked 2x PNG exports); pricing table also lists 3 daily check-in credits and 1 GB cloud storage.

Standard Monthly2026-10-21 ↗

USD 10.99/month

—

Monthly 100 credits/month + 5 daily check-in credits (up to 100 images or 10 videos/month), watermark-free exports, and 20 GB cloud storage. Rules [1]

Standard Annual2026-10-21 ↗

—

USD 71.88/year

Annual Billed annually at USD 71.88/year (USD 5.99/month equivalent); 100 credits/month + 5 daily check-in credits, watermark-free exports, and 20 GB cloud storage. Rules [2]

Premium Monthly2026-10-21 ↗

USD 19.99/month

—

Monthly 400 credits/month + 10 daily check-in credits (up to 400 images or 40 videos/month), Turbo speed, watermark-free exports, 50 GB cloud storage, and 15% off extra credits. Rules [1]

Premium Annual2026-10-21 ↗

—

USD 119.88/year

Annual Billed annually at USD 119.88/year (USD 9.99/month equivalent); 400 credits/month + 10 daily check-in credits, Turbo speed, watermark-free exports, and 50 GB cloud storage. Rules [2]

5 plan / billing-period records, all listed; — means that period is not recorded.

[1] Automatically renews monthly; cancel anytime

[2] Automatically renews annually; cancel anytime

The official pricing page lists Free Trial (5 free signup credits in the workspace where upscaling costs 2 credits/run, plus 3 daily check-in credits on the pricing table), Standard (USD 10.99/month or billed annually at USD 71.88/year, displayed as USD 5.99/month, with 100 credits/month + 5 daily check-in credits), and Premium (USD 19.99/month or billed annually at USD 119.88/year, displayed as USD 9.99/month, with 400 credits/month + 10 daily check-in credits and 15% off extra credits).

Cancellation, sources & verification record

Cancel auto-renewal anytime; the official pricing page states a 7-day money-back guarantee, and uploaded/generated files are automatically deleted within 7 days.

Media.io official monthly/annual pricing and comparison table ↗ · 2026-10-21 Capture 1 ↗ · Capture 2 ↗

Allowance & disclosure on test date

Credit cost ('Generate 2') is shown before submission, while the Watermark vs No Watermark (VIP) choice appears when clicking the download icon

Evaluation basis & full scores · Overall 75.1/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 66/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	20 / 35	Historical rubric scores withdrawn pending review of the new run. Used remaining 2 free credits for one Nano Banana 2 Lite image at 1K, 1:1, exact canonical prompt. Generated and downloaded native image with visible Media.io AI Gen watermark. Public visibility default ON shown in screenshot. No purchase. Correction: generation succeeded, but first download click produced no download event. Native export not yet saved; prior observation about downloaded native image was premature. Reopen result to finish download. Completed native download by selecting the Watermark option after opening the download menu. output.jpg is the actual delivered JPEG. Earlier click only opened that menu. Generation used 2 credits and exact 1:1 anchor prompt.	Evidence A →
Interaction restraint	17 / 30	Historical rubric scores withdrawn pending review of the new run. Used remaining 2 free credits for one Nano Banana 2 Lite image at 1K, 1:1, exact canonical prompt. Generated and downloaded native image with visible Media.io AI Gen watermark. Public visibility default ON shown in screenshot. No purchase. Correction: generation succeeded, but first download click produced no download event. Native export not yet saved; prior observation about downloaded native image was premature. Reopen result to finish download. Completed native download by selecting the Watermark option after opening the download menu. output.jpg is the actual delivered JPEG. Earlier click only opened that menu. Generation used 2 credits and exact 1:1 anchor prompt.	Evidence A →
Background & reputation	18 / 20	Historical rubric scores withdrawn pending review of the new run. Used remaining 2 free credits for one Nano Banana 2 Lite image at 1K, 1:1, exact canonical prompt. Generated and downloaded native image with visible Media.io AI Gen watermark. Public visibility default ON shown in screenshot. No purchase. Correction: generation succeeded, but first download click produced no download event. Native export not yet saved; prior observation about downloaded native image was premature. Reopen result to finish download. Completed native download by selecting the Watermark option after opening the download menu. output.jpg is the actual delivered JPEG. Earlier click only opened that menu. Generation used 2 credits and exact 1:1 anchor prompt.	Evidence A →
Engineering & privacy	11 / 15	Historical rubric scores withdrawn pending review of the new run. Used remaining 2 free credits for one Nano Banana 2 Lite image at 1K, 1:1, exact canonical prompt. Generated and downloaded native image with visible Media.io AI Gen watermark. Public visibility default ON shown in screenshot. No purchase. Correction: generation succeeded, but first download click produced no download event. Native export not yet saved; prior observation about downloaded native image was premature. Reopen result to finish download. Completed native download by selecting the Watermark option after opening the download menu. output.jpg is the actual delivered JPEG. Earlier click only opened that menu. Generation used 2 credits and exact 1:1 anchor prompt.	Evidence A →

Capability · 79/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	60 / 65	Used remaining 2 free credits for one Nano Banana 2 Lite image at 1K, 1:1, exact canonical prompt. Generated and downloaded native image with visible Media.io AI Gen watermark. Public visibility default ON shown in screenshot. No purchase. Correction: generation succeeded, but first download click produced no download event. Native export not yet saved; prior observation about downloaded native image was premature. Reopen result to finish download. Completed native download by selecting the Watermark option after opening the download menu. output.jpg is the actual delivered JPEG. Earlier click only opened that menu. Generation used 2 credits and exact 1:1 anchor prompt.	Evidence A →
Deliverable specifications	13 / 20	Native file obtained; see format/dimension inspection. Used remaining 2 free credits for one Nano Banana 2 Lite image at 1K, 1:1, exact canonical prompt. Generated and downloaded native image with visible Media.io AI Gen watermark. Public visibility default ON shown in screenshot. No purchase. Correction: generation succeeded, but first download click produced no download event. Native export not yet saved; prior observation about downloaded native image was premature. Reopen result to finish download. Completed native download by selecting the Watermark option after opening the download menu. output.jpg is the actual delivered JPEG. Earlier click only opened that menu. Generation used 2 credits and exact 1:1 anchor prompt.	Evidence A →
Processing speed	6 / 15	Network-observed server wait was not reliably isolated; speed and aggregate scores remain unknown.	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Raphael AI

raphael.app · Image generation

75/100

Seven-aspect task score

Login-free generation works, but the advertised model differs from the guest tier; free output is 640px, with about 85 seconds of waiting and a count error in this run.

Tested export 640×640 (0.5K Low WebP, 39.06% pixel retention vs 1024² baseline) · Historical description; output not attached

Historical test · 2026-10-16
Guest flow · Fixed English prompt · Historical page record; output not attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

Spelling was correct in this test; free output was 640px and the guest wait was about 85 seconds.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

Spelling was correct in this test; free output was 640px and the guest wait was about 85 seconds.

Original historical observation

Seedream 3.5 demonstrates strong typography ("REALTEST 2026" 25/25), but Raphael AI caps free users at 640×640 (0.5K Low) WebP and an ~85s queue despite its "100% Free · Powered by GPT Image 2.5" hero banner.

Suitable uses · historical assessment

Users who refuse to register an account and need unlimited, unwatermarked 640×640 concept drafts while tolerating an ~85s slow queue

Result & output evidence

No native output obtained; the preview does not establish final export quality.

Open result screenshot ↗

Full test record →

Evidence & scope →

Delivery capabilities

640×640 (0.5K Low WebP, 39.06% pixel retention vs 1024² baseline) · Historical description; output not attached · Check watermark record

Detailed facts & historical record

Recorded export specifications

640×640 (0.5K Low WebP, 39.06% pixel retention vs 1024² baseline) · Historical description; output not attached · Check watermark record

No native output is attached; check the evidence for specification and watermark claims.

Historical specification note · not an official maximum

640×640 (0.5K Low WebP · Free Guest Seedream 3.5; 2K/4K Require Paid Plan)

Watermark observation

Zero watermark (clean 640×640 WebP export)

Batch & API

Generates 2 variations per run; web workspace only

Recorded free-tier dimensions

Not verified

Attached file dimensions

Not verified

Evidence & scope →

Speed & reliability

About 85.5s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 85s · End-to-end reference: 85.5s
Date & fixture: 2026-10-16 · anchor_t2i_v1 · US-West (California, AS2914)
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

No login in this run; manual adjustment recorded

Detailed facts & historical record

Account requirements: No login in this test
Sign-in methods: No login required for 0.5K Low (640×640) guest generations
Manual adjustment: Manual erase / adjustment recorded in the historical flow
Interaction observations: Genuinely allows guests to generate and download 2 unwatermarked images without registering or logging in, though guests are routed into an ~85s slow queue ("In queue · Sign in for faster") with persistent "Upgrade for Fast Generation" upsell banners.

Evidence & scope →

Background & reputation

Raphael AI (raphael.app) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: Raphael AI (raphael.app)
Recorded origin: Independent AI Web Application
Domain history record: 2024–present
Background & reputation assessment: Popular high-traffic vertical AI image generator directory site (raphael.app, launched 2024).
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Uploaded for cloud processing；Claims deletion within 10 minutes; deletion not independently verified

Detailed facts & historical record

Processing location: Uploaded for cloud processing
Third-party tracking record: Not verified
Engineering & privacy assessment: Guest queue outputs 2× 640×640 unwatermarked WebP images (~43KB each) after an ~85.0s queue + inference wait.
Retention statement: Claims deletion within 10 minutes; deletion not independently verified

Evidence & scope →

Pricing & rules

Raphael AI free, monthly, yearly, and one-time credit pack plans (Free / Pro / Ultimate / Max / Credit Packs); USD · 2026-10-21

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free (Guest no-login)2026-10-21 ↗

Free

Free Unlimited no-login generations with Seedream 3.5 (0.5K Low · 640×640, slow queue); personal and non-commercial use only.

Pro Monthly2026-10-21 ↗

USD 20/month

—

Monthly 2,000 credits/month (up to 2,000 fast images or 200 basic videos) + unlimited Seedream 3.5, priority queue, no watermarks, full privacy, and commercial use.

Pro Yearly2026-10-21 ↗

—

USD 120/year

Annual Billed annually at USD 120/year (USD 10/month equivalent, regular USD 240/year); 2,000 credits/month + unlimited Seedream 3.5, no watermarks, and commercial use.

Ultimate Monthly2026-10-21 ↗

USD 40/month

—

Monthly 5,000 credits/month + unlimited Seedream 3.5 and Seedream 3.5 Pro, highest priority queue, no watermarks, full privacy, and commercial use.

Ultimate Yearly2026-10-21 ↗

—

USD 240/year

Annual Billed annually at USD 240/year (USD 20/month equivalent, regular USD 480/year); 5,000 credits/month + unlimited Seedream 3.5 / 3.5 Pro and commercial use.

Max Monthly2026-10-21 ↗

USD 80/month

—

Monthly 10,000 credits/month + unlimited Seedream 3.5, Seedream 3.5 Pro, MiniMax H3 Turbo, and MiniMax H3, highest priority queue, and commercial use.

Max Yearly2026-10-21 ↗

—

USD 480/year

Annual Billed annually at USD 480/year (USD 40/month equivalent, regular USD 960/year); 10,000 credits/month + unlimited generation on four core models and commercial use.

Credit Pack · 1,500 Credits2026-10-21 ↗

USD 29.9 · one-time payment

Usage / one-time

Usage / one-time 1,500 base credits (subscribers receive 2,250 credits with +50% bonus). Rules [1]

Credit Pack · 3,000 Credits2026-10-21 ↗

USD 49.9 · one-time payment

Usage / one-time

Usage / one-time 3,000 base credits (subscribers receive 4,500 credits with +50% bonus). Rules [1]

Credit Pack · 5,000 Credits2026-10-21 ↗

USD 69.9 · one-time payment

Usage / one-time

Usage / one-time 5,000 base credits (subscribers receive 7,500 credits with +50% bonus). Rules [1]

Credit Pack · 8,000 Credits2026-10-21 ↗

USD 99.9 · one-time payment

Usage / one-time

Usage / one-time 8,000 base credits (subscribers receive 12,000 credits with +50% bonus). Rules [1]

11 plan / billing-period records, all listed; — means that period is not recorded.

[1] Valid for 6 months

The official workspace and pricing page list Free (no-login unlimited Seedream 3.5 0.5K 640×640 in slow queue, personal/non-commercial use only), three monthly plans (Pro USD 20/mo, Ultimate USD 40/mo, Max USD 80/mo), three annual plans with explicit annual totals (Pro USD 120/yr, Ultimate USD 240/yr, Max USD 480/yr at 50% off), and four one-time credit packs valid for 6 months (1,500 / 3,000 / 5,000 / 8,000 base credits, with +50% bonus for subscribers).

Cancellation, sources & verification record

Cancel renewal anytime in the subscription management portal (benefits remain active until the end of the current billing period); FAQ states digital/virtual purchases are generally non-refundable.

Raphael AI official yearly, monthly, and credit pack pricing page ↗ · 2026-10-21 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗ · Capture 4 ↗

Allowance & disclosure on test date

Unlimited guest generations in the slow queue, restricted to Seedream 3.5 / Z-Image Turbo at 0.5K Low (640×640)

Model selector and resolution dropdown display Paid tags before clicking Generate, contradicting the hero banner's "100% Free · Powered by GPT Image 2.5" claim

Evaluation basis & full scores · Overall 74.7/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 74/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	25 / 35	Landing page hero badges (01_landing.png) prominently claim "100% Free · Powered by GPT Image 2.5 · No Login Required · Unlimited Generations"; however, the actual prompt box defaults guests to "Seedream 3.5" locked at "0.5K Low (640×640)", while "GPT Image 2.5", "Raphael 2.1", and "2K Medium / 4K High" resolutions carry Paid badges.	Evidence A →
Interaction restraint	23 / 30	Genuinely allows guests to generate and download 2 unwatermarked images without registering or logging in, though guests are routed into an ~85s slow queue ("In queue · Sign in for faster") with persistent "Upgrade for Fast Generation" upsell banners.	Evidence A →
Background & reputation	14 / 20	Popular high-traffic vertical AI image generator directory site (raphael.app, launched 2024).	Evidence C →
Engineering & privacy	12 / 15	Guest queue outputs 2× 640×640 unwatermarked WebP images (~43KB each) after an ~85.0s queue + inference wait.	Evidence A →

Capability · 75/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	58 / 65	Text Rendering (25/25: Exact 2-line "REALTEST 2026" carved onto the wooden sign with 100% spelling accuracy (25/25)) · Count & Color Binding (16/20: 2 blue ceramic mugs are exact, but the first image rendered 2 red apples instead of 3 (missing 1 apple; the 2nd image in the batch rendered 3 apples) (16.0/20)) · 5-Finger Hand & Structure (17/20: Right brass robotic hand gripping the wooden sign handle shows 4 fingers + 1 thumb with vintage brass patina, though overall clarity is limited by the 640×640 (0.5K Low) resolution cap (17.0/20))	Evidence B →
Deliverable specifications	14 / 20	Clean unwatermarked export (8/8); Export & Max deliverable spec: 640×640 (0.5K Low WebP · Free Guest Seedream 3.5; 2K/4K Require Paid Plan)	Evidence B →
Processing speed	3 / 15	Network-observed server wait: 85s (API mode: async_poll; end-to-end 85.5s shown for reference only, never penalized)	Evidence B →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Magic Studio

magicstudio.com · Image generation

44/100

Seven-aspect task score

The AI Art flow quickly generates and downloads 1024px JPEG images without a card gate; previews omit the exported bottom-right watermark, and text, object counts and hand constraints failed on this prompt.

Tested export JPEG · Output attached

Historical test · 2026-10-19
Account / tier unconfirmed · Fixed English prompt · Output attached

Seven-aspect rationale →

Verified native export & same-run screenshots

Inspect native image (JPEG · Output attached) ↗ Entry & rules ↗Result screen ↗Download & limits ↗

Use this capability ↗Tool profile →Full report →

Task quality

The brass robot and wooden table are recognizable, but the sign lacks the required text, two blue mugs are missing and the five-finger right hand is not met.

Detailed facts & historical record

Measured fixture results

Not verified

Fixture findings

The brass robot and wooden table are recognizable, but the sign lacks the required text, two blue mugs are missing and the five-finger right hand is not met.

Original historical observation

First native export from the AI Art path using the full standard prompt (1024x1024 JPEG, 0.93s median network-observed server wait): brass robot and dark wooden table are present (structure 10/20), two red apples and one yellow-green apple appear without the two blue mugs (count/color 7/20), and the wooden sign lacks legible REALTEST 2026 text (0/25); the exported file adds a bottom-right MagicStudio watermark absent in the preview.

Suitable uses · historical assessment

Not verified

Result & output evidence

Open native output ↗

Open result screenshot

Full test record →

Evidence & scope →

Delivery capabilities

JPEG · Output attached · Check watermark record

Detailed facts & historical record

Recorded export specifications: JPEG · Output attached · Check watermark record
Historical specification note · not an official maximum: Not verified
Watermark observation: Native JPEG has a bottom-right watermark absent in the inspected preview.
Batch & API: Not verified
Recorded free-tier dimensions: 1024 × 1024
Attached file dimensions: Not verified

Evidence & scope →

Speed & reliability

About 1.3s end to end in this run; reliability not retested

Detailed facts & historical record

Timing observation: Network-observed server wait: 0.93s · End-to-end reference: 1.3s
Date & fixture: 2026-10-19 · anchor_t2i_v1 · Not independently verified for this supplemental run
Repeated tests & reliability: No standardized repetition count, success rate or variability range is recorded.

Evidence & scope →

Interaction experience

Account requirement unconfirmed

Detailed facts & historical record

Account requirements: Not verified
Sign-in methods: Existing authorized test profile; no extra login in this path
Manual adjustment: Not verified
Interaction observations: Three serial downloads completed without a payment, card or additional login gate. Optional upgrade and below-fold promotion observed; watermark is not penalized again here.

Evidence & scope →

Background & reputation

aarzoo, Inc. (official declaration) · Long-term reputation not independently verified

Detailed facts & historical record

Service lifecycle notice: Not verified
Operator: aarzoo, Inc. (official declaration)
Recorded origin: Not verified
Domain history record: Not verified
Background & reputation assessment: Official terms name aarzoo, Inc. Independent reputation not checked: unknown midpoint 12.5/25 under the project specification.
External reputation record: Not verified

Evidence & scope →

Engineering & privacy

Processing location unconfirmed；Explicit retention period unconfirmed; inspect policy record

Detailed facts & historical record

Processing location: Processing location unconfirmed
Third-party tracking record: Not verified
Engineering & privacy assessment: HTTPS and three successful deliveries. Policy permits model improvement and describes deletion requests, targeting seven business days with possible remnants for 90 days. Automatic image retention duration not established; proposal 12/15.
Retention statement: Explicit retention period unconfirmed; inspect policy record

Evidence & scope →

Pricing & rules

Current Chinese web CNY Free and PRO 3/6/12-month plans · 2026-10-20

Plan · source date

Monthly / one-time

Annual bill / period

Quota & limits

Free2026-10-20 ↗

Free

Free Free without a usage-count limit; downloads up to 600px, one image at a time

Ads and Magic Studio logo

PRO · 3 months2026-10-20 ↗

CNY 149.99/3 months

Other period

Other period High-resolution downloads, no ads/branding, batch editing; access to all editing tools Rules [1]

PRO · 6 months2026-10-20 ↗

CNY 249.99/6 months

Other period

Other period High-resolution downloads, no ads/branding, batch editing; access to all editing tools Rules [1]

PRO · 12 months2026-10-20 ↗

—

CNY 239.99/year

Annual High-resolution downloads, no ads/branding, batch editing; access to all editing tools Rules [1]

4 plan / billing-period records, all listed; — means that period is not recorded.

[1] Current Chinese web CNY quote; 12-month tier shows a promotion. Other regional prices, renewal amount and paid pixel limit are not separately confirmed

Pricing now loads: 3/6/12-month totals are CNY 149.99/249.99/239.99. The current 12-month promotion is preserved without multiplying rounded monthly prices. Free use has no count limit but is capped at 600px with branding. Paid pixel limits, renewal amounts and other regional quotes are not stated in this component.

Verification has gaps; unknown prices and quotas are not inferred.

Cancellation, sources & verification record

FAQ states no refunds except where legally required; a separate renewal-cancellation flow was not verified

Official plans and usage terms ↗ · 2026-10-20 Capture 1 ↗

Loaded CNY subscription pricing and FAQs ↗ · 2026-10-20 Capture 1 ↗ · Capture 2 ↗ · Capture 3 ↗ · Capture 4 ↗

Allowance & disclosure on test date

Three free watermarked native exports; no unlimited-quota claim.

Not recorded

Evaluation basis & full scores · Overall 44.2/100

Overall = honesty & reputation × 30% + capability × 70%. Component maxima follow this report’s rubric. List scores are rounded for display; original precision is retained here.

Scoring version 2.3.1 · Test fixture anchor_t2i_v1

Honesty & reputation · 44.5/100 · 30%

Criterion	Score	Rationale	Basis
Pricing transparency	0 / 40	Review proposal 0/40 under the post-download restriction tier: inspected preview had no native-file watermark disclosure. Human confirmation is required; no new deception cap is inferred.	Evidence B →
Interaction restraint	20 / 20	Three serial downloads completed without a payment, card or additional login gate. Optional upgrade and below-fold promotion observed; watermark is not penalized again here.	Evidence B →
Background & reputation	12.5 / 25	Official terms name aarzoo, Inc. Independent reputation not checked: unknown midpoint 12.5/25 under the project specification.	Evidence C →
Engineering & privacy	12 / 15	HTTPS and three successful deliveries. Policy permits model improvement and describes deletion requests, targeting seven business days with possible remnants for 90 days. Automatic image retention duration not established; proposal 12/15.	Evidence C →

Capability · 44/100 · 70%

Criterion	Score	Rationale	Basis
Task quality	17 / 65	Single visual scoring proposal: text 0/25, count/color 7/20 (two red apples, one yellow-green apple, no required blue mugs), structure 10/20 (brass robot and wooden table present; sign not clearly held; five-finger right hand not met). Human calibration required.	Evidence A →
Deliverable specifications	12 / 20	First native JPEG: 1024x1024 = 12/12; visible watermark = 0/8. Paid maximum not tested.	Evidence A →
Processing speed	15 / 15	Verified network-observed server-wait median 0.93s over three serial runs = 15/15; not CPU/GPU-only time. Egress region unknown.	Evidence A →

Legacy scores may include advertised maximum capabilities. Read tested output separately from paid capabilities. Records without matching tiers and scope are not ranked together.

Tested tools withheld from Image generation ranking (5)

Factual blocker log · Unscored

These tools were tested using the same benchmark fixture for this topic, but hit a hard free-tier paywall / zero quota (B1) or backend error / missing standalone web entry (B2) before producing a verifiable free native export. They are listed here for transparency without speculative scores.

OpenArt (openart.ai) B1 · Hard paywall / 0 free export quota

Tested on 2026-11-02 using benchmark fixture anchor_t2i_v1: blocked by a free-tier paywall, credit gate, or backend error before native file export. Withheld from scoring and ranking under the report contract.

Tested on: 2026-11-02 · View product profile & pricing baseline →

Depositphotos (depositphotos.com) B1 · Hard paywall / 0 free export quota

Newly registered free accounts receive 0 generation credits; submitting the prompt redirects to the "Get Access" paid plan purchase page.

Tested on: 2026-10-22 · View product profile & pricing baseline →

Clipdrop (clipdrop.co) B1 · Hard paywall / 0 free export quota

Submitting the standard prompt after sign-in immediately displays "Image generation is for Pro" with zero free text-to-image quota.

Tested on: 2026-10-19 · View product profile & pricing baseline →

Picsart (picsart.com) B1 · Hard paywall / 0 free export quota

After Google sign-in, submitting the standard prompt in AI Playground leaves the canvas blank and prompts for paid credits or a Pro subscription.

Tested on: 2026-10-19 · View product profile & pricing baseline →

SnapEdit AI Background GeneratorGenerate BG (snapedit.app) B2 · Backend error / online entry unavailable

The public entry point is "Generate BG" (requiring an input photo for background replacement) rather than standalone text-to-image generation.

Tested on: 2026-10-19 · View product profile & pricing baseline →

Image generation selection FAQ

Which image generation tools work without login or watermarks?

Out of 28 assessed tools, 3 exported watermark-free files without requiring an account on their test dates, led by Vheer (89/100, 2026-10-17), Img2Go (87/100, 2026-10-19), MagicEraser.org (82/100, 2026-10-16). Use the "Quick guest use" preset above to filter all 3 options.

Which image generation tool scored highest on task quality and overall performance?

By overall task score, magnific.ai ranks first (95/100, tested 2026-11-02); by task quality score alone, magnific.ai leads (65/100), while Magic Studio recorded the fastest server wait (0.93s).

How should I choose between one-off guest use and commercial plans for image generation?

For a one-off file, start with the 3 guest watermark-free tools or pay-as-you-go credit packs to avoid recurring subscriptions. For team or client work, use the "Verified commercial plan" preset to inspect the 1 tools with sourced commercial rights.

What do these results cover?

Historical records cover specific fixtures and test paths. Reports show samples, dates, native files and limitations.

How are tools evaluated?

Evaluation covers quality, exports, speed, pricing, interaction, background and privacy. Scores and reasons are available in each report.

Image generation tools compared

Scan the candidates

Task quality

Delivery capabilities

Speed & reliability

Interaction experience

Background & reputation

Engineering & privacy

Pricing & rules

Honesty & reputation · 86/100 · 30%

Capability · 99/100 · 70%

Task quality

Delivery capabilities

Speed & reliability

Interaction experience

Background & reputation

Engineering & privacy

Pricing & rules

Honesty & reputation · 88/100 · 30%

Capability · 97/100 · 70%

Task quality

Delivery capabilities

Speed & reliability

Interaction experience

Background & reputation

Engineering & privacy

Pricing & rules

Honesty & reputation · 92/100 · 30%

Capability · 95/100 · 70%

Task quality

Delivery capabilities

Speed & reliability

Interaction experience

Background & reputation

Engineering & privacy

Pricing & rules

Honesty & reputation · 87/100 · 30%

Capability · 95.5/100 · 70%

Task quality

Delivery capabilities

Speed & reliability

Interaction experience

Background & reputation

Engineering & privacy

Pricing & rules

Honesty & reputation · 86/100 · 30%

Capability · 95/100 · 70%

Task quality

Delivery capabilities

Speed & reliability

Interaction experience

Background & reputation

Engineering & privacy

Pricing & rules

Honesty & reputation · 82/100 · 30%

Capability · 96.5/100 · 70%

Task quality

Delivery capabilities

Speed & reliability

Interaction experience

Background & reputation

Engineering & privacy

Pricing & rules

Honesty & reputation · 95/100 · 30%

Capability · 89/100 · 70%

Task quality

Delivery capabilities

Speed & reliability

Interaction experience

Background & reputation

Engineering & privacy

Pricing & rules

Honesty & reputation · 87/100 · 30%

Capability · 92/100 · 70%

Task quality

Delivery capabilities

Speed & reliability

Interaction experience

Background & reputation

Engineering & privacy