Test record & evaluation basis

Fish Audio · Audio transcription

Fish Audio offers guest-accessible online speech-to-text, achieving 100% keyword and verification token recall on the standard anchor with native TXT export.

Test date 2026-11-15Free tier + subscriptionNative output attachedMethod 2.3.1
1 · Date, environment & tier2026-11-15 · Guest flow

US-West (California, AS2914)

2 · Fixture & native export38.86s clean spoken English audio master (with verification token & numbers, 608 KB) (anchor_transcribe_v1)

TXT · Output attached · Open native file ↗

3 · Price checks vs. test scoreOfficial pricing checked 2026-11-12

Subsequent pricing and service checks are shown separately without altering historical scores.

4 · What one test cannot proveScoped to this fixture & path

One run does not establish all complex scenes, untested paid maxima or long-term uptime.

Plans & historical pricing

Site-wide TTS, Voice Design, and voice cloning plans on Fish Audio · 2026-11-12

Plan · source date
Monthly / one-time
Annual bill / period
Quota & limits
Free
Free
Free 8,000 credits per month (up to ~7 minutes audio)

Realistic AI voice synthesis, 3 public voice slots, enhanced voice cloning, no credit card required;500 chars/conv limit, personal non-commercial use only

Plus Plan (Monthly)2026-11-12 ↗
USD 15/month
—
Monthly 250,000 credits per month (up to ~200 minutes audio)

15,000 chars/conv, unlimited public + 10 private slots, priority queue, Voice Design, commercial rights

Plus Plan (Annual)2026-11-12 ↗
—
USD 66/year
Annual 250,000 credits/mo ($5.5/mo equivalent, $66/yr billed annually, save $114)

15,000 chars/conv, unlimited public + 10 private slots, priority queue, Voice Design, commercial rights

Pro Plan (Monthly)2026-11-12 ↗
USD 100/month
—
Monthly 2,000,000 credits per month (up to ~1,620 minutes audio)

30,000 chars/conv, 3 team seats, unlimited voice slots, 5 pro slots, 7-day refund guarantee

Pro Plan (Annual)2026-11-12 ↗
—
USD 450/year
Annual 2,000,000 credits/mo ($37.5/mo equivalent, $450/yr billed annually, save $750)

30,000 chars/conv, 3 team seats, unlimited voice slots, 5 pro slots, 7-day refund guarantee

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

Offers Free tier (8K credits/mo, ~7 mins); paid subscriptions include Plus ($15/mo, $66/yr introductory annual) and Pro ($100/mo, $450/yr introductory annual).

Cancellation, sources & verification record

Cancellation and refund terms unverified

Fish Audio Official Homepage & Studio ↗ · 2026-11-12 Capture 1 ↗

Fish Audio Official Pricing & Plans ↗ · 2026-11-12 Capture 1 ↗

Pricing recorded on the test date · 2026-11-15

Free guest tier; credit subscription from $15/mo

freemium_subscription

Historical context only; not current purchase guidance.

Delivery & access

Free delivery record
Native TXT text (0.5 KB, 64 words, with speaker labels)
Watermark
No watermark
Free allowance
Web tool supports direct single audio file conversion
Account requirements
No login required, full guest access
Maximum specification record
Native TXT text (0.5 KB, 64 words, with speaker labels)

May include advertised capabilities; this does not establish a tested paid tier.

Batch & API
Not recorded or not verified
Manual refinement
Not recorded or not verified

Tested result & sample

Fully retained RealTestTranscribeVerified2026 token and all financial figures with 100% recall

Fixture
anchor_transcribe_v1
Recorded environment
US-West (California, AS2914)
Network-observed server wait
48.22 s
End-to-end reference
48.22 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
Clearly shows supported formats prior to upload
Interaction
Not recorded or not verified
Entity record
Fish Audio (Hanabi AI Inc.)
Origin
Not recorded or not verified
Domain history record
Not recorded or not verified
Third-party tracking record
Google Analytics · Cloudflare
External information
Not recorded or not verified
Processing
Not recorded or not verified
Retention statement
Audio processed in real-time, not retained permanently

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.

Stage A · Benchmark fixture 38.86s clean spoken English audio master (with verification token & numbers, 608 KB)

anchor_transcribe_v1 · Inspect fixture & file inspector →

Stage B · Same-run screenshots 3 flow screenshots archived

Captures entry rules, result view and export disclosure.

Stage C · Native exported file TXT · Output attached

Open native exported file ↗

Stage D · Archived pricing sources 2 official source snapshots

Fish Audio Official Homepage & Studio (snapshot ↗) · Fish Audio Official Pricing & Plans (snapshot ↗)

Screenshots, native output and official claims support different conclusions. Evidence grades are not product safety certifications.

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_transcribe_v1

Honesty & reputation · 94/100 · 30%

CriterionScoreRationaleBasis
Pricing transparency34 / 35 Explanation unavailable Evidence A →
Interaction restraint28 / 30 Explanation unavailable Evidence A →
Background & reputation18 / 20 Explanation unavailable Evidence A →
Engineering & privacy14 / 15 Explanation unavailable Evidence A →

Capability · 83.8/100 · 70%

CriterionScoreRationaleBasis
Task quality57.8 / 65 Measured anchor metrics: word_accuracy_pct=79.4 · keyword_recall_pct=100 · token_preserved=true · duration_sec=38.9 · text_observed=Fully retained RealTestTranscribeVerified2026 token and all financial figures with 100% recall Evidence A →
Deliverable specifications16 / 20 Clean unwatermarked export (8/8); Export spec: Native TXT text (0.5 KB, 64 words, with speaker labels) Evidence A →
Processing speed10 / 15 Network-observed server wait: 48.22s (API mode: async_poll_verified; network upload 1.5s / end-to-end 48.22s 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

Prepare a correction or update →
Benchmark evidence preview

Open original ↗