🎤 AI-powered voice assistant with MCP integration - A fork of Whispo that transforms your voice into intelligent actions with advanced speech recognition, LLM processing, and Model Context Protocol (MCP) tool execution.

Click here to see v1 launch video on youtube

📥 Download Latest Release

Platform Support: macOS (Apple Silicon & Intel) with full MCP agent functionality. ⚠️ Windows/Linux: MCP tools not currently supported — see v0.2.2 for dictation-only builds.

Voice Recording:

1. Hold Ctrl (macOS/Linux) or Ctrl+/ (Windows) to start recording
2. Release to stop recording and transcribe
3. Text is automatically inserted into your active application

MCP Agent Mode (macOS only):

1. Hold Ctrl+Alt to start recording for agent mode
2. Release Ctrl+Alt to process with MCP tools
3. Watch real-time progress as the agent executes tools
4. Results are automatically inserted or displayed

Text Input:

• Ctrl+T (macOS/Linux) or Ctrl+Shift+T (Windows) for direct typing

git clone https://github.com/aj47/SpeakMCP.git && cd SpeakMCP
pnpm install && pnpm build-rs && pnpm dev

See DEVELOPMENT.md for full setup, build commands, troubleshooting, and architecture details.

AI Providers — Configure in settings:

• OpenAI, Groq, or Google Gemini API keys
• Model selection per provider
• Custom base URLs (optional)

MCP Servers — Add tools in mcpServers JSON format:

{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem", "/path"]
    }
  }
}

Keyboard Shortcuts:

We welcome contributions! Fork the repo, create a feature branch, and open a Pull Request.

💬 Get help on Discord | 🌐 More info at techfren.net

This project is licensed under the AGPL-3.0 License.

Built on Whispo • Powered by OpenAI, Anthropic, Groq, Google • MCP • Electron • React • Rust

Made with ❤️ by the SpeakMCP team