Happy slice of mcp pizzaMCP.pizza
Fire in da houseTop Tip:Paying $100+ per month for Perplexity, MidJourney, Runway, ChatGPT and other tools is crazy - get all your AI tools in one site starting at $15 per month with Galaxy AI Fire in da houseCheck it out free

basic-mcp

MCP.Pizza Chef: livekit-examples

basic-mcp is a voice assistant MCP server built on the LiveKit Agents framework, enabling voice-based AI interactions. It integrates MCP tools from external servers and supports speech-to-text via Deepgram, natural language processing with OpenAI's GPT-4o, text-to-speech, and voice activity detection using Silero. This server facilitates real-time, multimodal AI assistant workflows requiring voice input and external service integration.

Use This MCP server To

Enable voice-based AI assistant interactions with MCP tool integration Convert speech to text for natural language processing Use GPT-4o for advanced natural language understanding Perform text-to-speech synthesis for voice responses Detect voice activity to optimize interaction timing Integrate external MCP servers for extended functionality

README

LiveKit Agent with MCP Tools

A voice assistant application built using the LiveKit Agents framework, capable of using Multimodal Control Protocol (MCP) tools to interact with external services.

Features

  • Voice-based interaction with a helpful AI assistant
  • Integration with MCP tools from external servers
  • Speech-to-text using Deepgram
  • Natural language processing using OpenAI's GPT-4o
  • Text-to-speech using OpenAI
  • Voice activity detection using Silero

Prerequisites

  • Python 3.9+
  • API keys for OpenAI and Deepgram
  • MCP server endpoint

Installation

  1. Clone this repository:

    git clone https://github.com/livekit-examples/basic-mcp.git
cd basic-mcp

  2. Install the required packages:

    pip install -r requirements.txt

  3. Create a .env file with your API keys and configuration:

    OPENAI_API_KEY=your_openai_api_key
DEEPGRAM_API_KEY=your_deepgram_api_key
ZAPIER_MCP_URL=your_mcp_server_url

Usage

Run the agent with the LiveKit CLI:

python agent.py console

The agent will connect to the specified LiveKit room and start listening for voice commands.

Project Structure

  • agent.py: Main agent implementation and entrypoint
  • mcp_client/: Package for MCP server integration
    • server.py: MCP server connection handlers
    • agent_tools.py: Integration of MCP tools with LiveKit agents
    • util.py: Utility functions for MCP client

Acknowledgements

  • LiveKit for the underlying real-time communication infrastructure
  • OpenAI for GPT-4o and text-to-speech
  • Deepgram for speech-to-text
  • Silero for Voice Activity Detection

basic-mcp FAQ

How do I set up the basic-mcp server?
Clone the repository, install dependencies with pip, and configure API keys in a .env file.
What APIs does basic-mcp require?
It requires API keys for OpenAI, Deepgram, and an MCP server endpoint.
Which speech-to-text service does basic-mcp use?
It uses Deepgram for speech-to-text conversion.
How does basic-mcp handle natural language processing?
It uses OpenAI's GPT-4o model for natural language understanding and generation.
Can basic-mcp integrate with other MCP servers?
Yes, it supports integration with external MCP servers via MCP tools.
What voice activity detection technology is used?
basic-mcp uses Silero for voice activity detection to improve interaction responsiveness.
What Python version is required to run basic-mcp?
Python 3.9 or higher is required to run the server.
Is basic-mcp limited to OpenAI for NLP?
While it uses OpenAI's GPT-4o, it can be extended to support other LLM providers like Anthropic's Claude and Google's Gemini.