MCP - Typecast Documentation

One-Click Install

For Cursor and Replit, just click the button below:

One-click install for Cursor IDE

One-click install for Replit

Other MCP Clients

For other MCP-compatible clients, add this URL:

https://typecast.ai/docs/mcp

Claude Code

claude mcp add --transport http typecast-helper https://typecast.ai/docs/mcp

Windsurf

Open Windsurf Settings
Navigate to Cascade → MCP Servers
Click "Add Server" → "Add Remote MCP Server"
Enter URL: https://typecast.ai/docs/mcp

VS Code (Copilot)

Add to your .vscode/mcp.json:

{
  "servers": {
    "typecast-helper": {
      "url": "https://typecast.ai/docs/mcp"
    }
  }
}

That's it! You can now ask your AI assistant to help you integrate Typecast TTS into your projects.

What You Can Do

Once connected, your AI assistant gains knowledge about:

API Integration - Get code examples for any language
Voice Selection - Find the perfect voice for your use case
Best Practices - Learn optimal settings for different scenarios
Troubleshooting - Quick solutions for common issues

"Integrate Typecast TTS into my project."

Advanced: Automated TTS Generation

Need to generate audio files automatically? The hosted or self-hosted Typecast API MCP server lets your AI assistant directly call the Typecast API to create audio on demand, perfect for batch processing and automated workflows.

Both options also provide recommend_voices for natural-language voice search. Recommendation results contain only voice_id, voice_name, and score, so call get_voice or get_voices when your workflow needs metadata such as supported models, emotions, gender, age, or use cases.

What Makes This Different

Docs MCP	Hosted API MCP	Self-hosted API MCP
Provides knowledge and guidance	Calls Typecast API directly	Calls Typecast API directly
No API key required	No local installation	Runs on your computer
Best for integration help	Best for quick automation	Best for local files and audio playback

Prerequisites

Typecast API key (Get yours here)
uv package manager (self-hosted only)

Setup

Add the hosted Streamable HTTP endpoint to an MCP client that supports custom headers:

{
  "mcpServers": {
    "typecast": {
      "url": "https://typecast.ai/docs/mcp",
      "headers": {
        "X-API-KEY": "YOUR_API_KEY"
      }
    }
  }
}

Without the API key, the hosted server exposes only search_documentation. Authenticated requests unlock the Typecast API tools. Generated audio is returned as a private download URL that expires after one hour.

Troubleshooting

MCP server not appearing

Verify the configuration is correct
Restart your application completely
Check that uv is installed and available in your PATH

API key errors

Confirm your API key is set correctly in the config
Verify your key at Typecast API

Audio not playing (Linux)

Set XDG_RUNTIME_DIR environment variable
Check audio device: aplay -l

Resources

GitHub Repository

View source code for Self-hosted MCP Server

API Reference

Explore the Typecast API

Voice Library

Browse available voices

Python SDK

Build custom integrations

Control silence duration

The hosted Typecast API MCP server supports optional remove_silence_ms in standard, streaming, and timestamp TTS tools. Pass "remove_silence_ms": 300 as a tool argument. This API execution server is distinct from the documentation-search /docs/mcp; update self-hosted instances to a revision that includes this option.

remove_silence_ms specifies the silence duration to retain, not the amount to remove. Use an integer from 0 to 1000 ms. 0 removes detected silence; omission or null disables duration-based silence removal.

Standard, streaming, and timestamp TTS use output.remove_silence_ms; Compose uses segments[].output.remove_silence_ms on each tts segment. Returned timestamps align with the processed audio, and explicit pause segments are preserved.

Streaming's default leading-silence trimming is separate. Small values such as 0 can leave gaps between playable chunks; allow sufficient playback buffering and test with your content.