Text-to-speech API with ElevenLabs and MiniMax voices

The speech endpoint converts text to audio in thousands of voices from ElevenLabs, MiniMax, Microsoft Edge, Kokoro, Vbee and Fish Audio, and in voices you clone yourself. It also supports multi-speaker dialogue and pronunciation dictionaries.

Who it is for

Apps that narrate articles, lessons or videos
Products with voice responses
Podcast and audiobook workflows that need several speakers

How it works

Pick a voice from GET /v1/voices (ids carry a provider prefix).
POST the text; long texts are fine.
The finished task returns audio URLs, and optional SRT timings when with_transcript is true.

Endpoints

Method	Path	What it does
`POST`	`/v1/audio/speech`	Text to speech
`POST`	`/v1/audio/dialogue`	Multi-speaker dialogue
`GET`	`/v1/voices`	Search the voice library
`POST`	`/v1/dictionaries`	Pronunciation rules

Models

Pass one of these as model (or voice_id for voices). The full, current list with prices is at /v1/models.

elevenlabs_… voices
minimax_… voices
edge_… voices
kokoro_… voices
fishaudio_… voices
clone_… (your voices)

Example

curl https://omni-api.rcinc.app/v1/audio/speech \
  -H "Authorization: Bearer $OMNI_KEY" -H "content-type: application/json" \
  -d '{"input":"Welcome back to the show.","voice_id":"elevenlabs_TX3LPaxmHKxFdv7VOQHJ","speed":1}'

Pricing

ElevenLabs Multilingual v2 voices cost about $0.069 per 1,000 characters on Omni against ElevenLabs' $0.08; other voice libraries cost less. Each task shows its exact charge. See the pricing page for the full comparison.

Limits

Audio is generated asynchronously; short texts usually finish in seconds.
Cloned voices are private to the account that created them.
speed accepts 0.5 to 1.5.