Overview
eechat is a simple and easy-to-use locally deployed large model client, positioned as a local LLM chat tool focused on privacy, security, and performance. It is an open-source, cross-platform desktop application that supports Linux, macOS, and Windows, with full support for MCP (Model Context Protocol).
Key Capabilities
Privacy and Security
- All conversation data is stored locally, eliminating privacy risks.
- After setup is complete, it can be used offline to maximize data security.
Local Model Support
- AI models can be run directly on the device without sending data to external servers, giving users full control over their data.
API Integration
- You can connect your own AI model API or third-party services, with full control over data flow and model selection.
Content Rendering
- Supports perfect rendering of Markdown format, including code blocks, tables, and mathematical formulas.
- Automatically detects code blocks and performs multi-language syntax highlighting.
- Supports LaTeX, enabling rendering of complex mathematical equations and formulas.
MCP - Model Context Protocol
- Extends AI assistant capabilities through tool and service integration.
- One-click integration, management, and execution of multiple AI tools/services.
- The newly added MCP application supports visual configuration, can automatically parse README, and uses AI to assist in filling out configuration.
- Built-in runtime environments (bun, uv), with automatic detection and one-click installation support.
- MCP tools support hot-swapping, allowing instant add/remove/upgrade, and support multi-instance operation.
- Provides an intuitive management interface, allowing tools to be installed/removed without CLI operations.
Visual Configuration
- Easily add new MCP applications through form-based visual configuration.
- Supports automatically reading README or AI-assisted configuration filling, suitable for beginners.
Professional Configuration
- Built-in professional JSON configuration editor, supporting formatting, resetting, and directory navigation.
- Compatible with Claude Desktop, Cursor, and Cline configurations.
Integrated Runtime Environment
- Built-in runtime environments such as bun and uv, supporting automatic detection and one-click download, with no need for manual dependency configuration.
Chat Tool Integration
- All MCP tools can be called directly in the chat window, enabling seamless collaboration between AI and tools.
- Supports one-click switching of chat tools, with real-time status switching.
- Supports parallel processing of multiple tools, extendingconversation capabilities.
Supported Models
eechat supports a variety of local and API-based models, including:
- LLaVA: A multimodal model combining a vision encoder and Vicuna, used for vision and language understanding.
- Llama: A series of instruction-tuned image reasoning generation models, available in multiple sizes.
- DeepSeek: The first-generation reasoning model, with performance comparable to OpenAI-o1; DeepSeek's powerful language models can also be accessed via API.
- OpenAI: Connect to OpenAI's GPT models, including GPT-3.5 and GPT-4.
- Anthropic: Integrates Anthropic's Claude models for advanced reasoning.
- And more leading AI providers such as Gemini, Together, Perplexity, and others.
Pricing
- Free version: Suitable for ordinary users and testing, includes all core features, local model support, a 7-day free trial, and no credit card required.
- Lifetime version: Lifetime access, no subscription required. Includes all features of the free version, plus priority updates, advanced customization, premium support, multiple profiles, and early access to new features.
Download and Platforms
Download eechat on your preferred platform, free and open source:
- Windows 10/11 (64-bit)
- Apple Silicon Mac
- AppImage for most distributions
Open Source Project
eechat is fully open source. You can contribute code, report issues, or customize it as needed.
