multi-chat

Multi-Chat MCP Server is a Python-based framework that enables a single AI agent to operate across multiple chat providers simultaneously, with Google Chat being the production-ready integration. The server is architected with modular providers, so the AI can interact with Google Chat and, in the future, Slack and Microsoft Teams (already scaffolded for expansion). The core capability is a unified interface that lets the AI search history, share logs, summarize conversations, and propose actions across all connected platforms. To begin, install and run the server, then configure your environment so the AI agent can connect to Google Chat (and any additional providers you enable). Once running, you can issue tasks such as querying chat history, broadcasting messages, or coordinating cross-platform replies—all from a single AI context. Specific tools exposed by the Google Chat provider include actions like send, search, summarize, attach, and reply, enabling the agent to read conversations, insert responses, and attach artifacts as needed. As multi-provider support grows, you can extend the setup to include Slack and Teams, enabling the AI assistant to act as a participant across all configured channels at once.

Prerequisites:

• Python 3.8+ (recommended 3.11) and pip
• Git
• Optional: virtual environment tool (venv)

Install steps:

1. Clone the repository or install from the package manager if available:

   • git clone <repository-url>
   • cd multi-chat-mcp-server

2. Create and activate a virtual environment:

   • python -m venv venv
   • source venv/bin/activate

3. Install dependencies:

   • pip install -r requirements.txt

4. Install the MCP server package (if published to PyPI or via UV assert):

   • uvx install multi-chat-mcp-server or
   • pip install -e . # if running from source

5. Configure environment variables (as needed):

   • Set GOOGLE_CHAT_CREDENTIALS to the path of your Google Chat credentials JSON
   • Set LOG_LEVEL as desired (e.g., INFO or DEBUG)

6. Run the server:

   • uvx multi-chat-mcp-server
   • or python -m multi_chat_mcp_server

7. Verify the server is listening and the Google Chat provider is connected. Inspect logs for any authentication or API errors and adjust credentials accordingly.

Tips and common considerations:

• This project emphasizes local/on-premises operation for security and privacy; prefer on-prem or private LLM instances.
• The Google Chat provider is production-ready; Slack and Teams are scaffolds awaiting implementation, with the architecture in place to add them later.
• When running multiple providers, ensure your network and firewall settings allow needed API access for each platform.
• Use environment variables to securely provide credentials; do not commit sensitive keys to version control.
• Start with Google Chat only to validate the workflow, then add additional providers as needed.
• If using UV-based installation, keep the virtual environment activated during server runs to isolate dependencies.