ZenKnot

# ZenKnot > Multi-Tenant Intelligent Body Taming Platform · Easy Zen · Knowledge Graph Weaver [](LICENSE) [](https://fastapi.tiangolo.com/) [](https://vuejs.org/) [](https://www.langchain.com/) [](https://neo4j.com/) [](https://github.com/langchain-ai/deepagents) --- ## Introduction **ZenKnot** - A knot that ties thousands of tenants; a thought that solves complex graphs. The platform deeply integrates **LightRAG** knowledge base with **Neo4j** knowledge graph, built on **LangChain + Vue + FastAPI**, supporting multi-tenant isolation, DeepAgents deep intelligence, MinerU PDF analysis, and MCP protocol extension. Each tenant has an independent "Zen Garden" where intelligent agents retrieve, reason, and generate like flowing clouds. ## Core Features - **Multi-Tenant Zen Environment**: Tenant-level data isolation, independent knowledge graph namespace, and Agent sandbox - **LightRAG Intelligent Storage**: Vector + graph hybrid retrieval, semantic and structural dual recall - **Neo4j Knowledge Graph**: Knowledge graph storage and reasoning, relationships are the origin, and paths are the guidance - **DeepAgents Intelligent Agents**: LangChain-based deep intelligence agents, autonomous planning tool call - **MinerU Document Analysis**: PDF/document intelligent analysis, turning scriptures into readable laws - **MCP Protocol Extension**: Standard MCP protocol support, extending external tools and data sources at any time - **Vue User Interface**: Simple and elegant UI, tenant view, graph visualization, and conversation log - **FastAPI High-Performance API**: High-performance asynchronous API, automatically generating OpenAPI documentation ## Technology Stack | Domain | Technology Selection | |-----------|---------------------------------------| | Backend Framework | FastAPI + Uvicorn | | AI Orchestration | LangChain + DeepAgents | | Knowledge Graph | Neo4j + LightRAG | | Document Analysis | MinerU | | Protocol Extension | MCP (Model Context Protocol) | | Frontend Framework | Vue 3 + Vite + Pinia | | Database | Neo4j (main) + Vector Database (optional PGVector) | | Deployment | Docker Compose / Kubernetes | ## Quick Start ### Prerequisites - Python 3.10+ - Node.js 18+ - Neo4j 5.x (or use Docker to start) - (Optional) Docker Compose ### Clone Repository ```bash git clone https://github.com/your-org/ZenKnot.git cd ZenKnot ```