OpenDocuments
lastmod 2026-08-13created 2026-08-132 edits
An open-source RAG platform for searching scattered organizational documents in natural language
About
A self-hosted RAG platform built to solve the problem of team documents being scattered across GitHub, Notion, Drive, Confluence, and so on, which makes them hard to find.
Once you connect document sources, it parses, chunks, embeds, and indexes them, and when you ask a question in natural language it answers with sources.
- Connectors: GitHub, Notion, GDrive, S3, Confluence, Swagger, web crawler, Tavily web search
- File formats: 12 types including PDF, DOCX, XLSX, HTML, Jupyter, source code, and email
- LLM: can run locally with Ollama (no data leaks). Also supports OpenAI/Claude/Gemini/Grok
- Search: vector + keyword hybrid, reranking, hallucination guard
- Korean-English cross-lingual search
- Integrates with Claude Code/Cursor as an MCP server
- Web UI, 17 CLI commands, TypeScript SDK
- Plugin system (parsers/connectors/models/middleware)
- Supports Docker deployment