> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bland.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Migrate to Bland Speech

> Move an existing TTS integration to Bland's realtime WebSocket, HTTP, or OpenAI-compatible endpoint.

Choose the Bland endpoint that matches your application:

| Input shape                                                        | Use                                                  |
| ------------------------------------------------------------------ | ---------------------------------------------------- |
| LLM tokens or partial text, multi-turn conversation, interruptions | [`WSS /v2/tts/ws`](/api-v2/post/tts-ws)              |
| Existing OpenAI `/audio/speech` client or SDK                      | [`POST /v2/audio/speech`](/api-v2/post/audio-speech) |
| Complete text available up front, downloadable or batch audio      | [`POST /v2/tts`](/api-v2/post/tts)                   |

For new conversational applications, start with the [Realtime TTS Quickstart](/tts/realtime-quickstart). It keeps one warm WebSocket open, sends LLM tokens immediately, and lets Bland handle text buffering.

## Provider guides

<CardGroup cols={2}>
  <Card title="ElevenLabs" icon="arrow-right-arrow-left" href="/tts/migrations/elevenlabs">
    Replace stream-input messages, base64 audio, and manual flush behavior.
  </Card>

  <Card title="Cartesia" icon="arrow-right-arrow-left" href="/tts/migrations/cartesia">
    Map contexts and continuations to turns, preemption, and cancellation.
  </Card>

  <Card title="Deepgram" icon="arrow-right-arrow-left" href="/tts/migrations/deepgram">
    Map Speak, Flush, Clear, and Close to Bland turn messages.
  </Card>

  <Card title="PlayHT" icon="arrow-right-arrow-left" href="/tts/migrations/playht">
    Replace the generated socket URL and request-level speech commands.
  </Card>

  <Card title="Resemble AI" icon="arrow-right-arrow-left" href="/tts/migrations/resemble">
    Move from request IDs and audio\_end to Bland turn boundaries.
  </Card>

  <Card title="Google Cloud TTS" icon="arrow-right-arrow-left" href="/tts/migrations/google-cloud">
    Replace bidirectional gRPC configuration and audio responses.
  </Card>

  <Card title="Amazon Polly" icon="arrow-right-arrow-left" href="/tts/migrations/amazon-polly">
    Move synchronous synthesis or add token-level realtime streaming.
  </Card>
</CardGroup>

## Shared migration rules

All realtime migrations have the same core changes:

1. Select a Bland voice UUID from [List Voices](/api-v1/get/voices). Voice IDs are provider-specific and cannot be reused.
2. Open one socket per conversation and send `init` once.
3. Wait for `ready` before sending text.
4. Send only new text deltas under one `context_id` per turn.
5. Send `end_of_turn` after the LLM completes.
6. Read binary frames as raw PCM or mu-law audio.
7. On interruption, send a new context to preempt or send `cancel` without replacement text.
8. Drop unplayed local audio for preempted or cancelled turns.
9. Send `close` only after the final turn's `utterance_end`, then wait for `done`.

See [Realtime TTS Concepts](/tts/realtime-concepts) for the state model and playback responsibilities.

***

Docs for agents: [llms.txt](/llms.txt)