Vendor intelligence
Fish Audio
A developer speech API and creative voice workspace; paid TTS usage is metered in UTF-8 bytes.
Read the signals. Keep the scope.
A place in your stack.
Best for
- Developers prototyping expressive speech
- Narration pipelines that can meter input bytes precisely
Watch-outs
- UTF-8 bytes and characters are not interchangeable for multilingual scripts.
- Free API access is time-bound and has retention, commercial and service-level conditions.
Product specification
What the evidence says.
Unknown means we have not established it. It does not mean No. Compliance claims require plan and contract review.
White label
No supporting statement located in the bounded official corpus; confirm scope with vendor. Unknown is not No.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)Client tenancy
Workspace library and session variables do not establish reseller tenant isolation, reusable agent-template product or detailed role permissions.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)Client billing
No supporting statement located in the bounded official corpus; confirm scope with vendor. Unknown is not No.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)ISO 27001
Not established by the evidence reviewed. This does not mean unsupported.
No verified sourceBAA availability
No supporting statement located in the bounded official corpus; confirm scope with vendor. Unknown is not No.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)SIP
Import carrier-owned numbers through authenticated SIP trunks; carrier still bills PSTN legs.
Source ↗ (opens in new tab)Warm transfer
Phone-only handoff. Managed Twilio supports transfer; SIP cold requires REFER and warm requires termination. Warm briefing provides context.
Source ↗ (opens in new tab)Cold transfer
Phone-only handoff. Managed Twilio supports transfer; SIP cold requires REFER and warm requires termination. Warm briefing provides context.
Source ↗ (opens in new tab)Batch calling
Not established by the evidence reviewed. This does not mean unsupported.
No verified sourceMCP tools
Agent session tool timeline recognizes MCP-connected server calls; server setup and tools require configuration.
Source ↗ (opens in new tab)Webhooks
Fish executes an HTTPS request to a customer tool endpoint; CRM/calendar implementations remain customer-owned.
Source ↗ (opens in new tab)Calendar
No supporting statement located in the bounded official corpus; confirm scope with vendor. Unknown is not No.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)CRM
No supporting statement located in the bounded official corpus; confirm scope with vendor. Unknown is not No.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)Bring your carrier
Import carrier-owned numbers through authenticated SIP trunks; carrier still bills PSTN legs.
Source ↗ (opens in new tab)Bring your model
OpenAI-compatible custom endpoint receives conversation context and tool definitions; Fish handles remaining agent runtime. Not evidence of native support for every upstream provider.
Source ↗ (opens in new tab)Bring your voice
Voice and recognition selections are documented; no external provider credential/endpoint replacement path established for these stages.
Source ↗ (opens in new tab)HIPAA / BAA
Regulated-deployment sales language, payment-card processing and terms permissions do not establish HIPAA/PCI scope or a customer role matrix.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)SOC 2
No supporting statement located in the bounded official corpus; confirm scope with vendor. Unknown is not No.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)GDPR
No supporting statement located in the bounded official corpus; confirm scope with vendor. Unknown is not No.
Source ↗ (opens in new tab)Source ↗ (opens in new tab)Languages
Agents can select a speaking language and use Fish Library or own cloned voice. Not proof of automatic same-session language switching.
Source ↗ (opens in new tab)Voices
Not established by the evidence reviewed. This does not mean unsupported.
No verified sourceRealtime audio
Voice cloning
Architecture & connections
Understand the moving parts.
STT
Fish Audio transcribe models
LLM
Runtime provider options not established by current research
TTS
Fish Audio voice library / selected speech model
TELEPHONY
Unknown — provider details not established by reviewed evidence.
Integrations
Unknown — provider details not established by reviewed evidence.
Evidence you can inspect.
A documented feature, a supplier claim and an unanswered question are different kinds of evidence.
Evidence coverage details
- Strong-source resolved
- 61
- Vendor-claim only
- 0
- Conditional
- 19
- Other sourced outcomes
- 0
- Unknown after research
- 34
- Conflicting evidence
- 0
- Not applicable
- 0
- Not researched
- 0
Inspect detailed evidence, sources and research gaps
September 2026 remains the historical baseline. A new research date does not mean a feature launched that day. Unknown after research and not researched are separate states.
Official documentation establishes what the supplier documents. No paid calls, independent product benchmarks or compliance audits were performed in this research pass. Coverage measures research work, never product quality.
The operating decision.
Price structure
Component stack
V3.3 evidence review · TTS API reference; Voice Agents beta has separate minute-based billing
TTS uses UTF-8 bytes, ASR audio hours, Agents minutes plus carrier/model/add-ons. Multiple products must not share a fabricated normalized unit.
Human fallback
Unknown — no established fallback mode
No individually verified capability establishes this selection. Conditional support and vendor claims do not satisfy Strict Must-have.
Documented transfer is not a staffed human service. Verify who answers, when they answer and what happens if they do not.
Try it yourself
Unknown — no independently checked public demo entry is attached.
External signals
No dated external signal in this record. Missing evidence is not a negative review.
Underlying model context: the documented stack above does not establish an exact benchmark checkpoint or its deployed settings. Public model results cannot become a Fish Audio score.
Append-only price history
Every capture leaves a trace.
A new check creates a new snapshot. Deltas appear only for the same currency, unit, model and scope.
Tracked: $15 / 1m-utf8-bytes · Paid s2.1-pro TTS API
Tracked: $15 / 1m-utf8-bytes · V3.3 evidence review · TTS API reference; Voice Agents beta has separate minute-based billing
Tracked: $15 / 1m-utf8-bytes · V3.3 evidence review · TTS API reference; Voice Agents beta has separate minute-based billing
New research observations preserve earlier captures. “Evidence improved” records what we learned; it does not claim the market changed on that date. September 2026 is the historical baseline.
Price the actual workflow.
Complete cost cannot be established from this reference. Call mix, destination, transfer time, selected model/token usage and separate speech products prevent a universal complete monthly/minute ceiling. · Phone destination surcharge, $1.20/month managed number, cold $0.015/min or warm $0.025/min post-transfer, non-basic LLM tokens; BYO carrier separate
Carry the current source into your estimate, then enter the model, voice, carrier and fixed fees your configuration needs.
Alternatives by category
Other voice engines to inspect.
Alphabetical alternatives within the same category. An editorial starting point, not a performance ranking.
Historical capability and pricing research
CompareEvidence archive
Inspect the record.
Fish Audio API pricing and limitsreviewed
Paid s2.1-pro TTS is $15 per million UTF-8 bytes, not per million characters. s2.1-pro-free is $0. API billing has no subscription or monthly minimum. ASR is separately $0.36/audio hour. Starter concurrency is 5; higher tiers depend on total prepaid amount.
Inspect official source ↗Fish Audio capabilitiesreviewed
Official documentation covers REST/API and SDK access to TTS, ASR, voice cloning, voice management and real-time audio streaming. The free S2.1 model lacks TTFA and DPA guarantees.
Inspect official source ↗Fish Audio S2.1 free API conditionsreviewed
The updated free-API article says access currently runs through November 30, 2026 under fair use, without SLA or latency guarantees. Requests may be retained for model improvement. Businesses above $1M ARR should contact Fish before using the free model. The article lists 83 languages for S2.1 Pro.
Inspect official source ↗Fish Audio affiliate programreviewed
Fish Audio invites affiliate applications, issues a personal link after approval and lists monthly payouts. The public page does not state an exact commission rate. No approved VoiceAgent.best link is evidenced.
Inspect official source ↗Evidence describes documented scope, not a tested platform score. Unknown means a conclusion has not been established; API connectivity does not establish a native integration. Read our research methodology →
Application program verified; commission amount and API eligibility require confirmation. No tracking link on file.
Partner eligibility and commercial terms are Unknown in this review.