Voice MCP Agent

v1.0.0Communicationstable

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

voice-mcp-agentmcpai-integration
Share:
68
Stars
0
Downloads
0
Weekly
0/5

What is Voice MCP Agent?

Voice MCP Agent is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to voice assistant application built with the livekit agents framework, capable of using model context protocol (mcp) tools to interact with external services

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

This server falls under the Communication category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • A voice assistant application built with the LiveKit Agents

Use Cases

Build voice assistants using LiveKit Agents framework.
Integrate MCP tools into voice interactions.
Create conversational AI with external service access.
den-vasyliev

Maintainer

LicenseMIT
Languagepython
Versionv1.0.0
UpdatedMay 11, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx voice-mcp-agent

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Voice MCP Agent

Voice MCP Agent is a Python application built on the LiveKit Agents framework that creates a voice-controlled AI assistant capable of calling MCP tools to interact with external services. It uses OpenAI Whisper for speech-to-text, GPT-4o for reasoning, ElevenLabs for text-to-speech, and Silero for voice activity detection. The primary use case is voice-driven DevOps and SRE operations — users speak natural language commands and the agent translates them into Kubernetes management actions or other MCP tool calls. Note that the agent currently supports only HTTP/SSE MCP servers.

Prerequisites

  • Python 3.9 or later
  • An OpenAI API key (for Whisper STT and GPT-4o)
  • An ElevenLabs API key (for text-to-speech output)
  • A LiveKit server or LiveKit Cloud account
  • MCP servers accessible over HTTP/SSE (STDIO transport is not supported)
1

Clone the repository

Download the Voice MCP Agent source code.

git clone https://github.com/den-vasyliev/voice-mcp-agent.git
cd voice-mcp-agent
2

Create a virtual environment and install dependencies

Use the Makefile target to create the Python virtual environment and install all required packages.

make venv
source venv/bin/activate
make install
3

Set required environment variables

Export your API keys before running the agent. These are required for speech-to-text, reasoning, and text-to-speech.

export OPENAI_API_KEY="your-openai-api-key"
export ELEVEN_API_KEY="your-elevenlabs-api-key"
4

Configure your MCP servers

Edit `mcp_servers.yaml` to define which HTTP/SSE MCP servers the agent can call. Specify allowed tool patterns to restrict access.

# mcp_servers.yaml
servers:
  - name: k8s-mcp-server
    url: http://localhost:8092/sse
    allowed_tools:
      - list_*
      - describe_*
5

Start the voice agent

Launch the agent with `make run`. It connects to your LiveKit room and begins listening for voice input.

make run

Voice MCP Agent Examples

Client configuration

Voice MCP Agent is itself an MCP client, not a server you add to Claude Desktop. Configure it by editing `mcp_servers.yaml` to point at your HTTP/SSE MCP servers.

# mcp_servers.yaml
servers:
  - name: kubernetes
    url: http://localhost:8092/sse
    allowed_tools:
      - list_*
      - describe_*
      - scale_*
  - name: monitoring
    url: http://localhost:8093/sse
    allowed_tools:
      - get_metrics
      - alert_*

Prompts to try

Speak these commands to the voice agent. It will translate them into MCP tool calls against your configured servers.

- "Show me all resources in the default namespace"
- "List pods in the dev namespace"
- "Scale deployment my-app to 3 replicas in dev"
- "Describe pod my-app-123 in the default namespace"
- "What pods are currently in an error state?"

Troubleshooting Voice MCP Agent

Agent fails to call MCP tools — "transport not supported" error

Voice MCP Agent supports only HTTP/SSE MCP servers. Ensure the `url` in `mcp_servers.yaml` ends with `/sse` and the target MCP server has SSE transport enabled. STDIO-based servers are not supported.

No audio output after the agent responds

Verify `ELEVEN_API_KEY` is set correctly and the key has active character quota. Check that the ElevenLabs voice ID configured in the agent code exists in your ElevenLabs account.

Speech recognition is poor or misses commands

Ensure your microphone is selected as the default input device. Whisper performs best with clear audio — use a headset rather than built-in laptop microphone in noisy environments.

Frequently Asked Questions about Voice MCP Agent

What is Voice MCP Agent?

Voice MCP Agent is a Model Context Protocol (MCP) server that voice assistant application built with the livekit agents framework, capable of using model context protocol (mcp) tools to interact with external services It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Voice MCP Agent?

Follow the installation instructions on the Voice MCP Agent GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with Voice MCP Agent?

Voice MCP Agent works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Voice MCP Agent free to use?

Yes, Voice MCP Agent is open source and available under the MIT license. You can use it freely in both personal and commercial projects.

Browse More Communication MCP Servers

Explore all communication servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "voice-mcp-agent": { "command": "npx", "args": ["-y", "voice-mcp-agent"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Voice MCP Agent?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides