Generate Text-to-Speech

Synthesize speech from a script and get back a hosted audio URL. The returned url can be reused as an audio input in POST /v2/generate to lip sync the synthesized take. voiceId accepts an ElevenLabs voice id (discover via GET /v2/voices) or the id of a voice cloned via POST /v2/voices. Free-tier keys share the monthly ElevenLabs allowance; paid plans are billed per use.

A voiceId that does not resolve returns a 422 with errorCode voice_not_found. If a voice from your own ElevenLabs account stops resolving, the connected API key may be invalid — provider failures surface errorCode elevenlabs_api_key_invalid (reconnect it in Settings > Integrations) or elevenlabs_quota_exceeded (402) when your own ElevenLabs quota is exhausted. Temporary ElevenLabs provider failures return errorCode elevenlabs_service_unavailable: 504 for provider timeouts and 503 for other provider outages. Retry those requests with backoff and include the requestId if you contact support. A voice with an unresolved ElevenLabs restriction returns errorCode voice_verification_required or voice_blocked (both normally 422, or 400 when the pre-synthesis check was unavailable), or voice_access_denied (400). The required action depends on the code and on whether the voice is from your own connected ElevenLabs account — see the Error Handling guide.