Skip to main content
POST
cURL

Continue building

Write instructions

Choose intent, emotion, and delivery.

Convert text to speech

Generate audio without word timing.

Streaming timestamps

Stream audio with word timing.

Text to speech

Follow HTTP, SDK, and async examples.

Authorizations

xi-api-key
string
header
required

Breeze Developer API key.

Path Parameters

voice_id
string
required

Voice ID to use for speech generation. See List voices.

Query Parameters

output_format
enum<string>
default:pcm

Streaming audio encoding: pcm, mp3, or wav. Default: pcm. Audio is 24 kHz mono; PCM and WAV use 16-bit samples. See Output formats.

Available options:
pcm,
mp3,
wav
enable_logging
boolean | null

Body

application/json
text
string
required

Text to synthesize. Up to 1000 characters by default; accounts with an approved higher limit may send up to their configured limit, at most 2000 characters. See Audio tags.

Minimum string length: 1
model_id
string | null

Model ID for speech generation. When omitted, the server selects a model based on the account rollout, requested language, and endpoint capabilities. See List models.

Required string length: 1 - 120
language_code
string | null

ISO 639-1 two-letter language code supported by the selected model. See supported language codes.

Required string length: 2
Pattern: ^[A-Za-z]{2}$
instructions
string | null

Performance instructions, written in the same language as the input text. Leading and trailing whitespace is removed. See Expressive controls, Voice instruction prompting.

voice_settings
TtsVoiceSettingsPayload · object | null

Per-request voice settings. See Expressive controls.

Response

Streaming audio bytes. The history-item-id header identifies the persisted generation.

The response is of type file.