Summary

CatWiki is an enterprise-grade, full-stack AI knowledge base platform designed to provide robust content management, intelligent AI-powered Q&A, and a modern user experience. It targets businesses and developers seeking a comprehensive solution for organizing, accessing, and interacting with information, offering significant benefits through its advanced AI capabilities and flexible integration options.

How It Works

The platform employs a modern dual-client architecture, featuring a FastAPI backend and a Next.js frontend, both built with TypeScript and Pydantic for robust type safety. Its core AI functionality is powered by LangGraph, enabling sophisticated Agentic Retrieval Augmented Generation (RAG) with multi-turn autonomous retrieval and tool usage. This advanced approach allows the AI to dynamically query the knowledge base, continuously refining answers through iterative search and tool invocation, significantly enhancing accuracy and relevance. Data storage and vector embeddings are managed by PostgreSQL with the pgvector extension, while integrations with advanced document parsers like MinerU, Docling, and PaddleOCR ensure comprehensive content ingestion from various formats, including PDFs, Word documents, and images with OCR capabilities.

Quick Start & Requirements

• Primary install/run: Docker Compose for development (make dev-up), with a separate deploy/docker/ directory for production.
• Prerequisites: Docker (>= 20.10), Docker Compose (>= 2.0), Make. An OpenAI API Key is required for AI features.
• Links:
  • 官网: https://catwiki.ai
  • 演示站点: https://catwiki.cn/default/health
  • 管理后台: https://admin.catwiki.cn
  • 文档中心: https://docs.catwiki.cn

Highlighted Details

• Agentic RAG Workflow: AI autonomously utilizes the knowledge base via LangGraph for multi-turn, context-aware Q&A, demonstrating tool usage and iterative refinement.
• Extensive Multi-Platform Bot Integration: Offers deep support for DingTalk, WeChat, and Lark bots, alongside compatibility with OpenAI-protocol APIs, facilitating broad deployment scenarios.
• Advanced Document Parsing Engine: Integrates MinerU, Docling, and PaddleOCR for high-quality parsing of PDFs, Word documents, and images, including sophisticated layout analysis and OCR.
• Dual-Client Architecture: Provides distinct, well-defined interfaces for administration (content management, user roles, model configuration) and client-facing interaction (intelligent search, AI chat).
• Modern & Type-Safe Stack: Leverages FastAPI, Next.js 16, PostgreSQL, pgvector, Tailwind CSS, shadcn/ui, and uv package manager, ensuring a performant and maintainable codebase.

Maintenance & Community

The project shows active development with recent updates in February 2026 focusing on AI integration and bot support, and February 2024 on document parsing. It is maintained by the "CatWiki Team." Community feedback is managed via GitHub Issues, with contact information provided for business inquiries.

Licensing & Compatibility

The project is released under a CatWiki Open Source License, an adaptation of Apache 2.0. It permits free internal use but strictly prohibits the removal or modification of "CatWiki" branding in UI, console, or API responses. Crucially, it forbids providing commercial multi-tenant SaaS services without explicit written authorization from the CatWiki team.

Limitations & Caveats

The most significant limitation is the strict branding requirement and the prohibition of unauthorized commercial SaaS offerings, which may deter some businesses. While flexible for internal deployment, offering CatWiki as a managed service necessitates direct licensing. Full AI functionality relies on an external OpenAI API key.