ChatMCP is a cross-platform AI chat client designed to interact with Large Language Models (LLMs) via the Model Context Protocol (MCP). It aims to provide a unified interface for users to chat with various LLMs, manage chat history, and utilize features like SSE transport and theme switching.

How It Works

ChatMCP implements the Model Context Protocol (MCP) to abstract LLM interactions. This allows it to connect to different LLM providers (OpenAI, Claude, Ollama, DeepSeek) through a common interface. The client supports Server-Sent Events (SSE) for real-time responses and automatically selects available MCP servers.

Quick Start & Requirements

• Installation: Primarily distributed as a Flutter application. Installation instructions vary by OS (macOS, Windows, Linux, iOS, Android).
• Prerequisites:
  • Linux: libsqlite3-0, libsqlite3-dev.
  • Node.js (for uvx or npx installation of Node).
  • LLM API keys and endpoints configured in the application's settings.
• Configuration: MCP Server configuration is stored in mcp_server.json.
• Development: Requires flutter pub get and flutter run.
• Links: macOS Install, Windows Install, Linux Install, iOS TestFlight.

Highlighted Details

• Supports multiple LLM providers: OpenAI, Claude, Ollama, DeepSeek.
• Features Server-Sent Events (SSE) transport for real-time communication.
• Includes chat history and dark/light theme support.
• Planned features include an MCP Server Market and RAG integration.

Maintenance & Community

The project welcomes contributions for new features and bug fixes via its Issues tracker.

Licensing & Compatibility

• License: Apache License 2.0.
• Compatibility: Permissive license suitable for commercial use and integration with closed-source applications.

Limitations & Caveats

The project is actively under development, with several features marked as planned or incomplete, such as RAG and an MCP Server Market. The Linux installation requires manual installation of specific development libraries.