Connect desktop chat interface to MCP servers
A local-first AI chat app with a one-click GUI for installing, configuring, and invoking MCP tools - no command line required.
Why it matters
Enable users to interact with Model Context Protocol (MCP) servers through a unified desktop application that works across all major operating systems, providing a consistent chat interface for accessing MCP-connected tools and services.
Outcomes
What it gets done
Launch cross-platform desktop application on Linux, macOS, or Windows
Connect to multiple MCP servers from a single interface
Send chat messages to MCP-enabled services and tools
Receive and display responses from connected MCP servers
Install
Add it to your toolbox
Run in your project directory:
curl -fsSL https://spark.entire.vc/get/mcp-eechat | bash Overview
Eechat
A local-first AI chat application that manages MCP tools through a GUI - install, configure, and invoke them without touching a terminal - with a compatible JSON config editor for Claude Desktop, Cursor, and Cline. Use it when you want a private, offline-capable chat client with one-click MCP tool management instead of hand-editing configs or running a CLI to set up tool servers.
What it does
eechat is a local AI chat application designed for a secure, private, and efficient conversation experience, running fully on your own machine. Its core extension mechanism is MCP (Model Context Protocol) support: a management interface lets you view, install, start, stop, and remove MCP tools without any command-line work, and a form-based visual configurator can add new MCP applications by automatically reading a README or filling in configuration with AI assistance. For advanced users, eechat also ships a built-in JSON configuration editor (with formatting, reset, and quick navigation) that is compatible with Claude Desktop, Cursor, and Cline configuration formats, and it auto-detects and one-click-downloads runtimes like bun and uv so Node.js and Python-based MCP tools work without manual dependency setup. MCP tools can be added, removed, or upgraded at any time, support running multiple instances in parallel, and are invoked directly from the chat window itself.
When to use - and when NOT to
Use eechat when you want a local-first AI chat client that also manages MCP tool servers through a GUI - installing, configuring, and invoking them without touching a terminal, and without your conversation data leaving your machine (all data is stored locally, and the app can run offline after initial setup). It is not a bare MCP client library or SDK for embedding into another application - it is a standalone desktop chat app; if you need to integrate MCP into your own codebase rather than use a chat UI, this is the wrong layer.
Inputs and outputs
Inputs are chat messages plus whatever MCP tools you've installed and enabled through the management interface, configured either visually or via the JSON editor; model access is customizable to your own AI model API or third-party services such as OpenAI or Anthropic. Outputs are conversational responses with full Markdown rendering (code blocks, tables, math formulas) and code highlighting, plus whatever the invoked MCP tools return, surfaced directly in the chat window. Model behavior can be tuned per conversation (temperature, max output, and other parameters), and multiple independent chat sessions with saved history run side by side.
Integrations
Supports Claude Desktop, Cursor, and Cline configuration formats for MCP setup, and integrates with multiple AI model providers including OpenAI and Anthropic as well as custom local or remote models. Runtime dependencies for MCP tools (bun, uv) are auto-detected and installed by the app itself. Available as a pre-built package from the project's GitHub releases, or built from source:
# Clone repository
git clone https://github.com/Lucassssss/eechat.git
cd eechat
# Install dependencies
npm install
# Run in development mode
npm run dev
# Build application
npm run build
Runs on Windows 10/11 (64-bit), macOS 10.15+, and Linux (Ubuntu 18.04+, Debian 10+), needing at least 4GB RAM and 500MB disk space, with optional CUDA GPU support (8GB+) on Windows.
Who it's for
Users who want a private, local AI chat client with one-click MCP tool management instead of hand-editing config files or running a CLI - developers who want a GUI front end for Claude Desktop, Cursor, or Cline-style MCP configs, and anyone who wants built-in productivity features like document summarization, mind-map generation, and multi-language translation, or developer-focused features like code generation and debug assistance, in the same local chat interface rather than juggling separate tools.
Source README
๐ MCP feature is ready to rock!๐
eechat
๐ Powerful Local AI Chat Application - Secure, Efficient, Personalized
English | ็ฎไฝไธญๆ
Core Advantages โข Key Features โข Local Deployment โข Quick Start โข Tech Stack โข Contribute โข License
Core Advantages
eechat is an AI chat application focused on local deployment, providing users with a secure, private, and efficient AI conversation experience.
๐ MCP Support
MCP (Model Context Protocol) is a brand-new core extension capability of eechat, designed specifically for AI assistant scenarios. It greatly enhances the playability and professionalism of local AI. With MCP, you can easily integrate, manage, and run various AI tools and services with just one click, making your AI assistant infinitely expandable.
MCP Main Interface
Through the MCP management interface, you can easily view, install, start, stop, and remove various AI tools without command-line operations, significantly lowering the entry barrier.
MCP Visual Configuration
Provides a form-based visual way to add new MCP applications, supporting automatic README reading or AI-powered configuration filling. Even beginners can quickly integrate new tools.
MCP File Configuration
Built-in professional JSON configuration file editor, supporting formatting, reset, and quick navigation. Perfect for advanced users to batch manage and deeply customize. Fully compatible with Claude Desktop, Cursor, and Cline configurations.
Integrated Runtime Environment
Built-in automatic detection and one-click download for runtimes like bun and uv. No manual dependency setup required, ensuring Node.js/Python tools are ready to use out of the box.
Plugin Ecosystem & Hot-Swapping
MCP tools can be added, removed, or upgraded at any time, supporting multi-instance parallel running. Build your own exclusive AI toolbox.
Tool Invocation in Chat
All MCP tools can be called directly in the chat window, enabling seamless collaboration between AI and tools.
๐ Data Security & Privacy Protection
- Fully Local Storage: All conversation data is stored locally, eliminating privacy leak risks
- Offline Capability: No continuous internet connection required after initial setup
- API Customization: Connect to your own AI model API or third-party services, maintaining full data flow control
๐ฌ Superior Conversation Experience
- Multi-Session Management: Easily create and manage multiple independent conversations
- History Tracking: Automatically save all conversation history for review and continuation
- Perfect Markdown Support: Precise rendering of Markdown format, including code blocks, tables, and math formulas
- Smart Code Highlighting: Automatic code block detection and highlighting for multiple programming languages
๐จ Personalization Options
- Theme Switching: Built-in light and dark themes for different scenarios and preferences
- Prompt Management: Save and manage frequently used prompts to improve conversation efficiency
- Model Parameter Tuning: Flexibly adjust temperature, max output, and other parameters for optimal responses
๐ Powerful Extensibility
- Plugin Ecosystem: Support for functional extensions to meet specific scenario needs
- Multi-Model Integration: Easy integration with various AI model APIs like OpenAI, Anthropic, etc.
- Custom Model Support: Configure and use custom local or remote AI models
Key Features
Enhanced Intelligent Conversation
- Context Understanding: AI maintains context coherence in long conversations
- Multi-Turn Optimization: Enhanced interaction experience for complex queries
- Knowledge Base Integration: Connect custom knowledge bases for more accurate domain-specific responses
Developer-Friendly
- Code Generation & Explanation: Optimized code generation capabilities supporting multiple programming languages
- API Documentation Generation: Assist developers in quickly generating API documentation
- Debug Assistance: Help identify code issues and provide fix suggestions
Productivity Tools
- Document Summarization: Quickly summarize long document content
- Mind Map Generation: Transform complex concepts into structured mind maps
- Multi-Language Translation: Support high-quality translation between multiple languages
Local Deployment
eechat is designed for local deployment, ensuring your data security and optimal user experience.
System Requirements
- Windows 10/11 64-bit
- macOS 10.15+
- Linux (Ubuntu 18.04+, Debian 10+)
- Minimum 4GB RAM
- 500MB available disk space
- GPU with CUDA support (optional) 8GB+ (Windows)
Installation Methods
Download Pre-built Packages
Download the appropriate installation package for your system from the releases page:
Build from Source
# Clone repository
git clone https://github.com/Lucassssss/eechat.git
cd eechat
# Install dependencies
npm install
# Run in development mode
npm run dev
# Build application
npm run build```
FAQ
Common questions
Discussion
Questions & comments ยท 0
Sign In Sign in to leave a comment.