Audio

Generate speech from text

post

Generate speech from text using the Sesame (CSM 1B) TTS model via direct processing.

Available Models: sesame/csm-1b

Required Parameters

Parameter	Type	Description
text	string	Text to convert to speech (max 2000 characters)

Optional Parameters - Voice Selection

Parameter	Type	Default	Description
preset_voice	string	-	Pre-configured voice: Alice, Avery, Brock, Chloe, Ella, Emma, Grace, Karen, Kevin, Lucas, Matt, William
custom_voice	object	-	Clone a voice: {"url": "https://...", "context_text": "..."}
model	string	sesame/csm-1b	TTS model to use

Voice Selection Modes

1. Preset Voice (Recommended) - Use a pre-configured voice profile 2. Custom Voice Cloning - Clone any voice from an audio URL 3. Random Voice - Omit both for parametric synthesis with random voice

Note: Choose only ONE mode per request. Do not combine preset_voice with custom_voice.

Response

Returns binary WAV audio file (16-bit PCM, mono, 24kHz) with Content-Type: audio/wav

Async Mode

Set async: true to get a task_address immediately and poll for results. When complete, result contains media_key to fetch audio via /v2/audio/media/{media_key}.

Header parameters

x-api-keyany ofOptional

stringOptional

or

nullOptional

Body

application/json

Request schema for Text-to-Speech generation

textstring · min: 1 · max: 10000Required

Text to convert to speech

Example: The quick brown fox jumps over the lazy dog.

modelany ofOptional

TTS model. Available: 'sesame/csm-1b' (default)

Example: sesame/csm-1b

stringOptional

or

nullOptional

preset_voiceany ofOptional

Preset voice name. Use GET /v2/audio/tts/presets for available voices.

Example: Alice

stringOptional

or

nullOptional

custom_voiceany ofOptional

Custom voice cloning config (cannot use with preset_voice)

or

nullOptional

modeany ofOptional

Routing mode: 'auto' (intelligent), 'opengpu' (blockchain), 'direct' (low-latency)

Default: auto

string · enumOptionalPossible values:

or

nullOptional

asyncany ofOptional

Async mode: returns task_address immediately, poll /v2/tasks/{task_address} for result. Default: false (sync mode).

Default: false

booleanOptional

or

nullOptional

Responses

200

Successful Response

application/json

Responseany

202

Task accepted (async mode). Poll the poll_url for status.

application/json

422

Validation Error

application/json

post

/v2/audio/tts/sesame

HTTPcURLJavaScriptPython

POST /v2/audio/tts/sesame HTTP/1.1
Content-Type: application/json
Accept: */*
Content-Length: 51

{
  "text": "Hello, this is a test of text to speech."
}

Minimal (Recommended)With Preset VoiceWith Custom VoiceAll Options

No content

List available voice presets

get

List available TTS voice presets.

Returns a list of preset voice names that can be used with the preset_voice parameter. Available to both guests and authenticated users.

Header parameters

x-api-keyany ofOptional

stringOptional

or

nullOptional

Responses

200

Successful Response

application/json

422

Validation Error

application/json

get

/v2/audio/tts/presets

HTTPcURLJavaScriptPython

GET /v2/audio/tts/presets HTTP/1.1
Accept: */*

{
  "presets": [
    "text"
  ],
  "count": 1
}

Transcribe audio to text

post

Transcribe audio to text using Whisper large-v3 model.

Available Models: openai/whisper-large-v3

Required Parameters

Parameter	Type	Description
audio_url	string	URL to audio file (WAV, MP3, M4A, WEBM, FLAC, OGG)

Optional Parameters

Parameter	Type	Default	Description
language	string	-	Language hint for better accuracy (ISO 639-1 code, e.g., "en", "es", "fr")
model	string	openai/whisper-large-v3	ASR model to use
chunk_length_s	integer	30	Chunk length in seconds for processing long audio
batch_size	integer	24	Batch size for parallel chunk processing

Response

{
  "text": "Transcribed text content",
  "language": "en",
  "duration": 5.2
}

Supported Languages: 90+ languages including English, Spanish, French, German, Chinese, Japanese, etc.

Async Mode

Set async: true to get a task_address immediately and poll for results.

Header parameters

x-api-keyany ofOptional

stringOptional

or

nullOptional

Body

application/json

Request schema for Automatic Speech Recognition (Whisper).

Supports transcription (original language) and translation (to English).

audio_urlstringRequired

URL to audio file (WAV, MP3, M4A, WEBM, FLAC supported)

Example: https://example.com/audio.wav

modelany ofOptional

ASR model. Available: 'openai/whisper-large-v3' (default)

Example: openai/whisper-large-v3

stringOptional

or

nullOptional

taskany ofOptional

Task: 'transcribe' (original language) or 'translate' (to English)

Default: transcribe

string · enumOptionalPossible values:

or

nullOptional

languageany ofOptional

Language hint (ISO 639-1 code). Auto-detected if not specified.

Example: en

stringOptional

or

nullOptional

return_timestampsany ofOptional

Return timestamps: False (none), True (segment-level), 'word' (word-level)

Default: false

booleanOptional

or

const: wordOptional

or

nullOptional

temperatureany ofOptional

Sampling temperature (0.0 = deterministic)

number · max: 1Optional

or

nullOptional

chunk_length_sany ofOptional

Process audio in chunks of this length in seconds (default: 30)

integer · min: 1 · max: 60Optional

or

nullOptional

batch_sizeany ofOptional

Batch size for processing chunks (default: 16)

integer · min: 1 · max: 32Optional

or

nullOptional

modeany ofOptional

Routing mode: 'auto' or 'direct' (ASR is direct-only)

Default: auto

string · enumOptionalPossible values:

or

nullOptional

asyncany ofOptional

Return task_address immediately, poll /v2/tasks/{task_address} for result. Default: false (synchronous).

Default: false

booleanOptional

or

nullOptional

Responses

200

Successful Response

application/json

202

Task accepted (async mode). Poll the poll_url for status.

application/json

422

Validation Error

application/json

post

/v2/audio/asr/whisper

HTTPcURLJavaScriptPython

POST /v2/audio/asr/whisper HTTP/1.1
Content-Type: application/json
Accept: */*
Content-Length: 77

{
  "audio_url": "https://www.soundhelix.com/examples/mp3/SoundHelix-Song-1.mp3"
}

Minimal (Recommended)With Language HintAll Options

{
  "text": "text",
  "language": "text",
  "duration": 1,
  "task_address": "text",
  "mode": "text",
  "done": true,
  "ANY_ADDITIONAL_PROPERTY": "anything"
}