Voice API economics

Find the cheapest TTS API for delivered speech

Compare text-to-speech providers by their documented character or UTF-8-byte meter and accepted audio, while keeping model, voice, language, latency, and rights fixed.

One page · one buyer job

cheapest TTS API

Checked Aug 22, 2026

Estimate the cost of approved synthesized speech and shortlist a TTS route that meets voice, language, latency, and policy requirements.

Meter
Characters, bytes, tokens, seconds, or credits
Voice scope
Model, language, cloned voice, and streaming mode
Quality
Pronunciation, prosody, stability, and rerender rate
Operations
Latency, concurrency, caching, and audio delivery

TTS text-unit cost calculator

Copy the current per-1,000 rate for the exact voice model and use that provider's documented character or UTF-8-byte unit throughout the estimate.

Use characters for a character-metered route or encoded UTF-8 bytes for a byte-metered route; never mix them in one result.
Convert the exact current provider rate to 1,000 of the same native unit before entering it.
Extra billable text units from rejected or revised speech generations.
Reset example

Planning result

Review the example inputs, replace the rate with the exact current provider rate, and calculate a workload. APIDir does not send these values to a provider.

Native meters, not a false universal ranking

Voice API pricing: Fish Audio and ElevenLabs

Character, UTF-8 byte, audio-hour, request, concurrency, language, and commercial-use boundaries are kept separate. Recheck the official page before purchase.

Audit the evidence method →

Pay-as-you-go API pricing with no subscription fee or monthly minimum

Fish Audio API pricing

  • s2.1-pro, s2-pro, and s1: $15 per million UTF-8 bytes
  • transcribe-1: $0.36 per audio hour, rounded up to the nearest second
  • voice-design-1: $0.01 per successful API request

Languages: Fish Audio documents 80+ languages for S2-Pro with automatic language detection; S1 documents 13 languages.

Concurrency: The documented prepaid thresholds unlock 5 concurrent requests below $100 paid, 15 at $100 or more, and 50 at $1,000 or more. Enterprise limits are custom.

Commercial use: Fish Audio's terms license paid-service users for commercial use subject to the agreement, applicable law, input rights, and content restrictions. This is not a substitute for reviewing the current terms for a specific voice or dataset.

Measurement rule: Count the encoded UTF-8 bytes sent after templates, pronunciation markup, and localization. Do not treat one character as one byte for multilingual text.

API product rates shown separately from monthly plan allowances and taxes

ElevenLabs API pricing

  • Flash, Turbo, and v3 Conversational TTS: $0.05 per 1,000 characters
  • v2 Multilingual and v3 TTS: $0.10 per 1,000 characters
  • Scribe v2: $0.22 per audio hour; Scribe v2 Realtime: $0.39 per audio hour

Languages: ElevenLabs documents 70+ languages for v3, 29 for Multilingual v2, and 32 for Flash/Turbo. Model choice also changes request-size and latency boundaries.

Concurrency: Published TTS concurrency varies by plan and model family: for example, Starter documents 6 concurrent Flash/Turbo requests or 3 other-model requests, while Pro documents 20 or 10. Enterprise can request custom limits.

Commercial use: Paid plans include commercial rights for compliant output when the customer holds the necessary input rights; free-plan output is non-commercial, and beta services have separate restrictions. Voice cloning still requires authorization and policy compliance.

Measurement rule: Use billed characters from the exact TTS model and keep TTS, transcription, speech-engine, dubbing, and audio-processing meters separate.

Untracked official: Official link: affiliate tracking is not active for this destination, so APIDir does not receive referral credit from this click. Editorial evidence, rankings, price order, and corrections are computed independently of commission.

ElevenLabs referral disclosure

Published program: 22% of qualifying Starter, Creator, Pro, and Scale payments for the referred subscriber's first 12 months; 11% for Business; no enterprise commission.

The current program uses PartnerStack onboarding and a unique tracked link with a published 90-day attribution window. The referred customer must register and maintain a qualifying non-free subscription under the current terms. The published terms say commissions are paid at the end of the third month after they are earned and can be adjusted for refunds or cancellations.

APIDir's outbound-click event contains only the provider, page surface, and link mode. A qualifying conversion is reported by the provider's partner system and is not joined to email addresses or calculator inputs.

Official program page · Program terms · checked Aug 22, 2026

Make the native billing unit explicit

TTS pricing can count characters, UTF-8 bytes, tokens, audio duration, or credits. ElevenLabs publishes character rates, while Fish Audio publishes rates per million UTF-8 bytes. A multilingual character may occupy several bytes, so converting both providers to a fake character rate would distort the comparison. Count the final encoded request after templates, pronunciation markup, and localization, then keep that native unit through the calculation.

The calculator adds a rerender allowance because pronunciation errors, clipped audio, style changes, and editorial rejection create more billable work. For a conversational product, separate pre-generated prompts from real-time responses. Cached fixed audio and live generation have different cost, latency, and availability profiles.

A voice shortlist needs a listening protocol

Use the same scripts across names, numbers, dates, acronyms, code-switching, emotional passages, long-form narration, and noisy conversational turns. Record the exact model, voice ID, language, stability or style settings, output format, sample rate, and streaming mode. Blind reviewers should score pronunciation, naturalness, speaker consistency, pacing, and task suitability.

Quality and price are not the only gates. Confirm commercial rights, voice consent, cloning restrictions, data retention, regional processing, content policy, latency percentiles, concurrency, rate limits, streaming behavior, and model-version stability. A low character rate cannot compensate for a voice that fails the product's language or compliance requirement.

  • Track billed provider-native text units from usage, not only source-document length.
  • Separate text-to-speech, speech-to-text, voice agents, and audio-processing meters.
  • Cache reusable approved audio where the license and product experience allow it.
  • Create a fallback voice and a visible degraded state for provider outages.

Keep the voice integration replaceable

Map product concepts such as voice, language, speed, and output format into an internal request, then adapt it to each provider. Do not scatter provider voice IDs and unsupported parameters across the application. The adapter should return a stable job ID, usage fields, provider request ID, output metadata, and a typed error category.

Retry only requests that are documented as safe, cap retries, and check whether a timed-out job was accepted before submitting again. Log identifiers and error class without storing text or audio unnecessarily. This architecture makes a provider switch testable, reversible, and deletable.

Common comparison mistakes

  • Treating a UTF-8 byte as one character or assuming every provider counts SSML the same way.
  • Comparing a fast conversational model with a long-form quality model as one offer.
  • Ignoring rerenders caused by pronunciation and editorial approval.
  • Mixing TTS, STT, voice-agent, and audio-processing prices.
  • Hardcoding provider voice IDs throughout product code.

Questions developers ask

What is the cheapest TTS API?

Among routes that pass your language, quality, latency, rights, and policy gates, compare the measured cost using each route's documented native text unit and your rerender rate.

Do SSML tags count toward TTS pricing?

The rule varies. Read the exact provider documentation and use billed usage; do not assume visible text length equals billable characters.

Should speech-to-text be compared on this page?

No. STT is normally billed by audio duration and has a different accuracy job. Keep it as a separate scenario even when one provider sells both products.

First-party boundaries

Official sources to recheck

These links define meters and product boundaries. APIDir does not freeze their current rates inside the planning calculator.

How APIDir handles evidence →

APIDir Delta · controlled rollout

Reserve source-linked price research.

Confirm your address to join the rollout. Editorial delivery starts only after sender, suppression, and evidence checks are active.