SuperAPI AI Directory

chatmcp

ChatMCP

Open website

About this tool

ChatMCP is an AI chat client that implements the Model Context Protocol (MCP), with the project hosted on GitHub.

In terms of purpose, ChatMCP is used for conversations with MCP-enabled servers, allowing users to connect different data sources for chatting.

In terms of main capabilities, the information shows that it supports the following features: cross-platform, covering macOS, Windows, Linux, iOS, Android, and Web; support for both stdio and SSE MCP transports; support for automatically installing MCP Servers and automatically selecting MCP Servers; compatibility with LLM models such as OpenAI, Claude, OLLama, and DeepSeek; dark/light themes; data can be synchronized between ChatMCP applications on the same local area network.

In terms of usage, users need to configure the LLM API Key and endpoint address on the settings page, then install an MCP Server from the MCP Server page, after which they can chat with the MCP Server. Before use, you need to ensure that uvx or npx is installed on the system. The project provides an MCP Server marketplace, and users can also submit a Pull Request for their own MCP Server to the mcp_server_market repository to join the marketplace. Developing or running ChatMCP requires installing Flutter. The repository enforces code formatting requirements through the .githooks/pre-commit hook, which runs dart format . on every commit, and a format check also runs in CI; commits with incorrect formatting will be rejected locally and CI will fail.

In terms of applicable scenarios, it is suitable for users who need to use the MCP protocol to connect data sources for AI conversations across multiple operating systems.

In terms of limitations, the information does not provide more details about pricing, cost, specific feature effectiveness, or complete usage steps. Running AppImage/DEB on Linux requires installing specific runtime libraries, for example AppImage requires libfuse2, and in addition requires GTK 3, graphics/EGL-related libraries, X11 tools, and SQLite, etc., with optional mesa-vulkan-drivers and mesa-utils for diagnostics. The Web version runs entirely in the browser, and chat history and settings are stored locally.storage.