How to Connect AnyToSpeech to Cursor (MCP Setup)
Remote MCP · sign-in or integration key

How to Connect AnyToSpeech to Cursor

Generate voiceovers for demo videos, narrate release notes or transcribe user interviews without leaving your editor. AnyToSpeech is a remote MCP server, so there's nothing to install.

Start free See the steps
Setup

Set up AnyToSpeech in Cursor

The easiest way is the Add to Cursor button on your AnyToSpeech Integrations page. To set it up by hand, follow these steps.

  1. Open mcp.json

    Use ~/.cursor/mcp.json for every project, or .cursor/mcp.json for one repo.

  2. Add the server

    Add an anytospeech entry with the server URL (see the config below). No headers or keys are needed: you sign in instead.

    https://anytospeech.com/api/mcp

  3. Connect

    Open Cursor Settings → Tools & MCP. AnyToSpeech shows Needs authentication. Click Connect, sign in to AnyToSpeech in your browser and approve.

  4. Ask the agent

    In Agent chat, ask for audio. The AnyToSpeech tools appear in the MCP tools list and the agent calls them when needed.

~/.cursor/mcp.json. For CI or headless use, add "headers": { "Authorization": "Bearer <integration key>" } with an integration key from your Integrations page.

{
  "mcpServers": {
    "anytospeech": {
      "url": "https://anytospeech.com/api/mcp"
    }
  }
}
Prompts

Prompts to try

“Write and generate a 60-second voiceover for this feature from the README.”
“Narrate CHANGELOG.md for this release as an MP3.”
“Transcribe interview.mp3 and summarise the key user complaints.”
“Generate English and Spanish versions of this onboarding script.”
FAQ

Frequently asked questions

Sign-in or integration key?

Use sign-in for day-to-day use: Cursor handles it and keeps you connected. Use an integration key from your Integrations page for CI pipelines or the headless CLI. Disconnect signed-in apps under Connected apps and revoke keys under Integration keys.

Does it work with other MCP clients?

Yes. Any client that supports remote Streamable HTTP MCP servers can use the same URL, including VS Code, Windsurf, Claude Code, Codex CLI and Meta's Muse Code.

Where do generated files go?

Tools return download links to the audio. Ask the agent to save the file into your project if you want it in the repo.

Keep going

Related guides

MCP server overview Every tool, scope and auth option.
Connect to Claude Works on Claude Free too.
Connect to ChatGPT Developer mode setup.

Add voice to your workflow

Create a free AnyToSpeech account and use the one-click Add to Cursor button.

Create a free account MCP server details