> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.resemble.ai/voice-generation/text-to-speech/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.resemble.ai/_mcp/server.

# Text-to-Speech

Turn text into natural, production-ready speech. Resemble supports multiple synthesis modes tuned for different latency and integration needs.

New and upgraded voices use **Resemble Ultra** by default. Provide a `voice_uuid` with your request; the API automatically uses that voice's model version.

## Synthesis Modes

### [Synchronous](./text-to-speech/synchronous)

Request-based synthesis that returns a complete audio file in a single response.

Best suited for:

* Alerts and notifications
* Short-form content
* Workflows that require the entire clip before progressing

### [Streaming over HTTP](./text-to-speech/streaming-http)

Receive audio chunks progressively via chunked HTTP responses.

Best suited for:

* Longer scripts
* Progressive playback experiences
* Reducing perceived latency without persistent sockets

### [Streaming over WebSocket](./text-to-speech/streaming-websocket)

Maintain a WebSocket to receive the lowest-latency audio stream with per-chunk metadata.

Best suited for:

* Conversational agents
* Interactive assistants
* Real-time media experiences where milliseconds matter (Business plan and above)

## Next Steps

1. Generate an API token from the dashboard.
2. Pick the synthesis mode that fits your UX.
3. Follow the dedicated page for request/response formats and implementation tips.