What is an MCP server for voice?
Short answer: A service an AI assistant can call, through the Model Context Protocol, to speak text or look up audio instead of you writing glue code.
In plain terms
MCP is a standard way for an AI assistant (Claude, Cursor and others) to use outside tools. A voice MCP server gives it tools such as "speak this text" and "search the phrase library".
Speakvora's server
| Tool | What it does |
|---|---|
| speak | Turns text into speech and returns an audio link |
| list_voices | Lists live voices and languages |
| search_phrases | Searches the checked phrase library so the assistant can reuse an existing line |
Endpoint: https://speakvora.com/api/mcp (streamable HTTP), auth with your API key in x-api-key. It is also listed in the official MCP Registry as io.github.eyeguide-dev/speakvora.
Why it is useful
- An assistant can prototype spoken features without you writing the integration.
- Searching first reuses lines that already exist as audio.
Set it up: connect to Claude Code and Cursor · MCP server page