chatmcp

ChatMCP is a cross-platform AI chat client that comes with an integrated MCP (MCP) server experience suitable for local and LAN usage. It is a Flutter-based desktop and mobile application that lets you chat with an AI model, while providing built-in MCP server endpoints (stdio and SSE variants) to enable other tools or clients on your network to connect to and reuse the chat context. The app stores data locally, including chat history, settings, and MCP server configurations, following platform conventions. To begin, install the app on your platform, then open the Settings or MCP Server page to configure your LLM API key and endpoint, which the client will use to generate responses. The Data Storage section describes where your data is kept and how to reset it if needed. You can also use the Data Sync feature to share chat data with other devices on the same network.

Once running, you can interact with the built-in MCP servers (stdio and SSE variants) for simple and streaming interaction patterns. The stdio MCP server is geared toward traditional request/response interactions, while the SSE MCP server supports server-sent events for real-time streaming updates. In addition to chat capabilities, the app offers data storage and preferences management, allowing you to customize themes, storage locations, and where your chat history is kept. The documentation also notes Linux-specific runtime considerations and optional dependencies if you plan to run the Linux package in certain environments.

Prerequisites:

• A supported platform (macOS, Windows, Linux, iOS, Android, or Web) and the ability to run Flutter desktop or Web apps as applicable.
• Flutter SDK installed on your development machine if you plan to build from source. See the official Flutter docs for installation instructions.

Installation steps (official releases):

1. Download the ChatMCP app from the official release page for your platform (macOS, Windows, Linux, iOS TestFlight, Android, or Web).
2. Install or unzip the app following platform-specific guidelines.
3. Run the application:
   • Desktop (macOS/Windows/Linux): Launch the ChatMCP executable or via your app launcher.
   • Web: Open the hosted web version in your browser.
4. Open the Settings or MCP Server page inside the app to configure your LLM API key and endpoint. This is required for the MCP server to forward requests to your chosen LLM provider.
5. If you are developing locally from source, ensure Flutter is installed, then run:

   # Clone repository
   git clone https://github.com/daodao97/chatmcp.git
   cd chatmcp
   
   # Install dependencies
   flutter pub get
   
   # Run on macOS/Linux (desktop)
   flutter run -d macos
   # Or run for Linux desktop
   flutter run -d linux
   

6. To build a release for Linux, use:

   flutter build linux
   

Prerequisites note: For Linux, ensure your environment has Flutter desktop support enabled and the system meets any runtime requirements listed in the project documentation.

Tips and considerations:

• The app stores data locally. See the Data Storage section in the README for exact paths per platform and how to reset data if needed.
• You can use the Data Sync feature to share data across devices on the same LAN.
• When configuring LLM integration, provide a valid API key and endpoint. The app will use these to generate responses via the MCP servers.
• Linux users may need platform-specific runtime libraries for AppImage/DEB deployments (as documented in the Linux notes). Ensure dependencies such as libfuse2, libgtk-3-0, and graphics libraries are installed if you’re deploying a Linux package.
• If you plan to run a local MCP server, you’ll typically interact with the stdio or SSE endpoints exposed by the app. Ensure your network allows the necessary connections and that your firewall permits the local traffic.
• The project emphasizes code formatting and pre-commit hooks for contributors; if you build from source, consider enabling the repository’s formatting hooks to maintain consistency.