Changelog

2026-09-14

Added

• POST /v1/voices/add Instant Voice Clone (multipart files or JSON sample_url). Returns {voice_id, requires_verification: false} while extract runs; poll GET /v1/voices/{voice_id} for status.
• DELETE /v1/voices/{voice_id} soft-delete.
• Voice objects include status (processing | ready | failed) and optional workspace_id / error.
• GET /v1/voices lists all known clones in the bound library (workspace-aware). Query status=ready for usable-only.
• OAuth oat_ tokens are accepted on Speech. New scope tts:write for add/delete.
• Hosted MCP: braiv_list_voices, braiv_get_voice, braiv_add_voice, braiv_text_to_speech, braiv_delete_voice.

Current limitations

• WAV output only. No streaming, MP3/PCM, or Professional Voice Clone.

Beta launch

Initial developer documentation for the Braiv Text to Speech API.

Added

• ElevenLabs-compatible POST /v1/text-to-speech/{voice_id} endpoint.
• GET /v1/voices and GET /v1/voices/{voice_id} for voice clone discovery.
• GET /v1/models with model id braiv-speech-v1.
• API key authentication with xi-api-key, bearer token, and query parameter fallbacks.
• Plan gating for Pro and Business users.

Current limitations

• WAV output only.
• No public streaming endpoint yet.
• English-only docs for v1.