AgentChat

Overview

AgentChat is a comprehensive AI dialogue system with MCP server integration. It features a three-layer memory architecture, including short-term context retention, historical summarization, and long-term user preference tracking. The platform supports Human-In-The-Loop (HITL) mechanisms for dialog-based MCP server generation from OpenAPI specifications. Built with FastAPI backend and Vue 3 frontend, it offers multi-agent collaboration, RAG technology for knowledge retrieval, and supports various vector databases like Milvus and ChromaDB.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• query_weather

  查询实时天气信息和预报

  Try:查询北京的天气Try:明天上海会下雨吗？
• generate_image

  AI驱动的图像生成服务

  Try:生成一只卡通猫的图片Try:创建一张科幻城市的概念图
• create_mcp_server

  基于人机协同方式从OpenAPI信息生成MCP Server

  Try:根据这个API文档帮我创建MCP服务器Try:协助我配置这个REST API为MCP工具

Note: Tool names inferred from the feature descriptions and screenshots. The README mentions support for tools and MCP servers but doesn't provide explicit tool names or a dedicated tools section.

Comparable tools

Installation

# 1. Clone the repository
git clone https://github.com/Shy2593666979/AgentChat.git
cd AgentChat

# 2. Edit configuration file
vim docker/docker_config.yaml

# 3. Start the service
cd docker
docker-compose up --build -d

# Backend setup
cd src/backend
pip install -r requirements.txt
uv sync  # Recommended

# Frontend setup
cd src/frontend
npm install
npm run dev

{
  "mcpServers": {
    "agentchat": {
      "command": "python",
      "args": ["path/to/agentchat/mcp_server.py"]
    }
  }
}

FAQ

Does AgentChat support multiple LLM providers?

Yes, AgentChat supports multiple AI models and provides a unified interface to switch between different LLM providers.

How do I add custom tools to AgentChat?

You can add custom tools by uploading Swagger/OpenAPI specifications through the platform's interface, which will automatically generate MCP server bindings for the new tools.

Compare AgentChat with