MagicEraser.org (MagicEraser 6.0 Text-to-Image) · Image generation
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.
US-West (California, AS2914)
JPEG · Output attached · Open native file ↗
Subsequent pricing and service checks are shown separately without altering historical scores.
One run does not establish all complex scenes, untested paid maxima or long-term uptime.
Plans & historical pricing
Web subscriptions, displayed top-ups and enterprise quote; USD · 2026-10-20
11 plan / billing-period records, all listed; — means that period is not recorded.
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 ↗
Pricing recorded on the test date · 2026-10-16
10 Free Credits for guests (1 credit/image); Pro plans from ~$6.9/mo
Freemium + credit subscription
Historical context only; not current purchase guidance.
Delivery & access
- Free delivery record
- 1000×1000 (unwatermarked JPG)
- Watermark
- No watermark
- Free allowance
- 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)
- Account requirements
- No login required (guests receive 10 Free Credits automatically)
- Maximum specification record
- 1000×1000 (Guest no-login unwatermarked JPG export)
May include advertised capabilities; this does not establish a tested paid tier.
- Batch & API
- Supports batchMode toggle
- Manual refinement
- Directly switch to Erase or Fine Edit on the generated image
Tested result & sample
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.
- Fixture
- anchor_t2i_v1
- Recorded environment
- US-West (California, AS2914)
- Network-observed server wait
- 8.96 s
- End-to-end reference
- 9.36 s
Network wait includes round trips, queuing and polling error, not pure inference time. Local timings depend on the device; repetition counts are not standardized in legacy records.
Experience, rules & trust
- Disclosure
- Displays "Credits Required: 1 Credit" directly above the Generate button
- Interaction
- One-click Download button exports the unwatermarked JPG and allows instant switching to Erase, Enhance, or Upscaler in the same workspace.
- Entity record
- MagicEraser.org / LinoCut AI
- Origin
- US / APAC Independent AI Team
- Domain history record
- 2023–2026
- Third-party tracking record
- Google Analytics · Aliyun SLS Web Tracking
- External information
- One of the very few AI Image Generators that allows guests to generate and download 1000×1000 unwatermarked images without logging in.
- Processing
- Cloud async generation queue (me_single_photo_editor_v1)
- Retention statement
- Signed OSS URLs expire in ~4 hours
A recorded statement does not verify deletion or all data-processing behavior.
Verifiable audit trail & evidence
Four SHA-256-verified stages link the benchmark fixture, same-run interface screenshots, native exported artifact and archived pricing sources.
anchor_t2i_v1 · Inspect fixture & file inspector →
Captures entry rules, result view and export disclosure.
Official plans and usage terms (snapshot ↗) · Large top-up package selections (snapshot ↗)
Screenshots, native output and official claims support different conclusions. Evidence grades are not product safety certifications.
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.
Revisions & corrections
- 2026-10-16
Initial real-machine benchmark using anchor_t2i_v1 5-constraint prompt.