signal-mcp-client

MCP.Pizza Chef: piebro

Signal MCP Client is a Model Context Protocol client that integrates with the Signal messaging platform to send and receive messages securely. It enables real-time communication within MCP workflows using Signal's encrypted messaging infrastructure, supporting multi-step reasoning and interaction with LLMs. This client is designed for users who want to incorporate Signal-based messaging into their AI-enhanced workflows across platforms.

Use This MCP client To

Send and receive MCP messages via Signal's encrypted network Integrate Signal messaging into AI-driven chat workflows Enable secure, real-time communication for LLM agents Use Signal as a transport layer for MCP client-server interactions Facilitate multi-step reasoning with Signal message context Deploy AI chatbots that communicate over Signal Connect Signal conversations to external MCP servers and tools

README

Signal MCP Client

An MCP (Model Context Protocol) client that uses Signal for sending and receiving messages.

Setup and start the Signal Chat Bot

These Instructions are for Ubuntu Linux. With some minor modification this should also work on other Linux Distros, Mac or Windows. I recommend to use an extra phone number for the bot, so you don't have to use your own.

  1. Install uv and podman:
    sudo apt install podman
    curl -LsSf https://astral.sh/uv/install.sh | sh
  2. Start the Signal CLI Rest Server container:
    mkdir -p $HOME/.local/share/signal-api
    podman run \
        --name signal-cli-api \
        --replace \
        -p 8080:8080 \
        -v $HOME/.local/share/signal-api:/home/.local/share/signal-cli \
        -e 'MODE=json-rpc' \
        docker.io/bbernhard/signal-cli-rest-api:latest-dev
  3. Connect the signal-cli-rest-api container to your signal account by opening this link and scanning the QR code:
    http://localhost:8080/v1/qrcodelink?device_name=signal-api
    
  4. Create a config.json with your mcp servers. For example for testing you can use:
    cat << EOF > config.json
    {
        "servers": [{
            "name": "echo-mcp-server-for-testing",
            "command": "uvx",
            "args": ["echo-mcp-server-for-testing"],
            "env": {"SECRET_KEY": "123456789"}
        }]
    }
    EOF
  5. Create a session directory for saving the message history, images, videos and settings for each user and start the MCP client:
    export ANTHROPIC_API_KEY='your-key'
    export SIGNAL_PHONE_NUMBER='+1234567890'
    export FAL_KEY='your-key' # optional for transcribing voice messages
    
    uvx signal-mcp-client \
        --config config.json \
        --session-save-dir /absolute/path/to/session/dir \
        --available-models claude-3-7-sonnet-latest claude-3-5-haiku-latest \
        --default-model-name claude-3-7-sonnet-latest \
        --default-system-prompt "" \
        --default-llm-chat-message-context-limit 50

Adding MCP Server

Add the MCP server in the config.json file.

Here are some example servers I use:

  • vlc-mcp-server: An MCP Server to play and control local movies using VLC.
  • fal-ai-mcp-server: An MCP Server to use the fal.ai APIs to generate images and videos.

Contributing

Contributions to this project are welcome. Feel free to report bugs, suggest ideas, or create merge requests.

Development

Running from source

  1. Clone the repo git clone git@github.com:piebro/signal-mcp-client.git.
  2. Go into the root dir cd signal-mcp-client.
  3. Run the MCP client:
    export ANTHROPIC_API_KEY='your-key'
    export SIGNAL_PHONE_NUMBER='+1234567890'
    export FAL_API_KEY='your-key' # optional for transcribing voice messages
    # additional optional environment variables:
    export SIGNAL_WS_BASE_URL="ws://localhost:8080"
    export SIGNAL_HTTP_BASE_URL="http://localhost:8080" 
    export CLIENT_LOG_LEVEL="DEBUG"
    export SERVER_LOG_LEVEL="DEBUG"
    # you can also use a .env file in the root directory to set the environment variables
    
    uv run signal_mcp_client/main.py \
        --config config.json \
        --session-save-dir /absolute/path/to/session/dir \
        --available-models claude-3-7-sonnet-latest claude-3-5-haiku-latest \
        --default-model-name claude-3-7-sonnet-latest \
        --default-system-prompt "" \
        --default-llm-chat-message-context-limit 50

Formatting and Linting

The code is formatted and linted with ruff:

uv run ruff format
uv run ruff check --fix

Building with uv

Build the package using uv:

uv build

Releasing a New Version

To release a new version of the package to PyPI, create and push a new Git tag:

git tag v0.2.0
git push origin v0.2.0

Running as a Systemd Service

To ensure the Signal REST API and the MCP Client run automatically on boot and restart if they fail, you can set them up as systemd user services. User services run under your specific user account.

This setup assumes that you have completed the setup steps and your project is located at /home/$USER/signal-mcp-client.

  1. Enable User Lingering to keep your user session active after logging out.

    sudo loginctl enable-linger $USER
  2. Create Systemd Service Directory

    mkdir -p /home/$USER/.config/systemd/user/
  3. Create Service File for Signal REST API

    cat << EOF > "/home/$USER/.config/systemd/user/signal-cli-rest-api.service"
    [Unit]
    Description=Signal CLI REST API Container
    After=network.target
    Wants=network-online.target
    
    [Service]
    Environment="XDG_RUNTIME_DIR=/run/user/%U"
    Environment="DBUS_SESSION_BUS_ADDRESS=unix:path=%t/bus"
    SyslogIdentifier=signal-cli-rest-api
    Restart=on-failure
    RestartSec=30
    
    ExecStartPre=-/usr/bin/podman stop signal-cli-api
    ExecStartPre=-/usr/bin/podman rm signal-cli-api
    
    ExecStart=/usr/bin/podman run --name signal-cli-api \\
        -p 127.0.0.1:8080:8080 \\
        --replace \\
        -v /home/$USER/.local/share/signal-api:/home/.local/share/signal-cli \\
        -e MODE=json-rpc \\
        docker.io/bbernhard/signal-cli-rest-api:latest
    
    ExecStop=/usr/bin/podman stop signal-cli-api
    
    [Install]
    WantedBy=default.target
    EOF
  4. Create Service File for Signal MCP Client

    cat << EOF > "/home/$USER/.config/systemd/user/signal-mcp-client.service"
    [Unit]
    Description=Signal MCP Client Application
    After=network.target signal-cli-rest-api.service
    Wants=signal-cli-rest-api.service
    
    [Service]
    Environment="ANTHROPIC_API_KEY=your-key" 
    Environment="SIGNAL_PHONE_NUMBER=+1234567890"
    # Environment="FAL_KEY=your-key"
    # Environment="SIGNAL_WS_BASE_URL=ws://127.0.0.1:8080"
    # Environment="SIGNAL_HTTP_BASE_URL=http://127.0.0.1:8080" 
    # Environment="CLIENT_LOG_LEVEL=INFO"
    # Environment="SERVER_LOG_LEVEL=INFO"
    SyslogIdentifier=signal-mcp-client
    
    Restart=on-failure
    RestartSec=30
    
    ExecStart=/home/$USER/.local/bin/uvx signal-mcp-client \
        --config config.json \
        --session-save-dir /home/$USER/sessions \
        --available-models claude-3-7-sonnet-latest claude-3-5-haiku-latest \
        --default-model-name claude-3-7-sonnet-latest \
        --default-system-prompt "" \
        --default-llm-chat-message-context-limit 50
    
    [Install]
    WantedBy=default.target
    EOF
  5. Enable and Start the Services

    systemctl --user daemon-reload
    
    systemctl --user enable signal-cli-rest-api.service
    systemctl --user enable signal-mcp-client.service
    
    systemctl --user start signal-cli-rest-api.service
    systemctl --user start signal-mcp-client.service
  6. Check Service Status and Logs

    systemctl --user status signal-cli-rest-api.service
    systemctl --user status signal-mcp-client.service
    
    journalctl --user -u signal-cli-rest-api.service -f
    journalctl --user -u signal-mcp-client.service -f

License

This project is licensed under the MIT License. See the LICENSE file for details.

signal-mcp-client FAQ

How do I set up the Signal MCP Client?
Follow the GitHub instructions for installation on Linux, Mac, or Windows, including installing dependencies like uv and podman.
Can I use my personal Signal number for the client?
It is recommended to use a separate phone number for the bot to avoid mixing personal and bot messages.
Is communication via Signal MCP Client secure?
Yes, it leverages Signal's end-to-end encryption for all message exchanges.
What platforms does the Signal MCP Client support?
Primarily Linux, with potential support for Mac and Windows with minor modifications.
Can the Signal MCP Client interact with multiple MCP servers?
Yes, it can send and receive messages to/from various MCP servers integrated with Signal.
Does the client support multimedia messages?
The client supports sending and receiving images as shown in the GitHub screenshots.
How does the Signal MCP Client handle message context?
It maintains conversation context to enable multi-step reasoning and interaction with LLMs.
Which LLM providers can be used with this client?
It is provider-agnostic and can work with OpenAI, Anthropic Claude, Google Gemini, and others.