OpenDocuments

Open source RAG tool for AI document search — connect GitHub, Notion, Google Drive and ask questions with cited answers.

Install

npm install -g opendocuments

Quick Start

opendocuments init     # Auto-detects Ollama, pulls models, validates API keys
opendocuments start    # Web UI at http://localhost:3000
opendocuments index ./docs
opendocuments ask "How does authentication work?"

What It Does

OpenDocuments connects your scattered documents and answers questions with AI:

• 10+ data sources — GitHub, Notion, Google Drive, Confluence, S3, Swagger, web pages
• 12+ file formats — PDF, DOCX, Excel, HTML, Jupyter, code, email, PowerPoint
• Local or cloud AI — Ollama (data stays local) or OpenAI, Claude, Gemini, Grok
• Korean + English — cross-lingual search finds docs regardless of language
• MCP server — works with Claude Code, Cursor, and any MCP-compatible AI tool

CLI Commands

# Ask questions
opendocuments ask "query"              # Single question
opendocuments ask                      # Interactive REPL
opendocuments search "keyword" --top 5 # Vector search (no LLM)

# Documents
opendocuments index ./docs --watch     # Index + auto-reindex
opendocuments document list            # List indexed docs
opendocuments document delete <id>     # Soft-delete

# Connectors
opendocuments connector sync           # Sync all sources
opendocuments connector status         # Check sync status

# Server
opendocuments start                    # HTTP + Web UI
opendocuments start --mcp-only         # MCP server (stdio)
opendocuments stop                     # Stop server
opendocuments doctor                   # Health diagnostics

# Plugins
opendocuments plugin create my-parser --type parser
opendocuments plugin list
opendocuments plugin publish

# Admin
opendocuments auth create-key --name "bot" --role member
opendocuments export --output ./backup

Configuration

opendocuments init generates opendocuments.config.ts:

import { defineConfig } from 'opendocuments-core'

export default defineConfig({
  workspace: 'my-team',
  mode: 'personal',

  model: {
    provider: 'ollama',        // or 'openai', 'anthropic', 'google', 'grok'
    llm: 'qwen2.5:14b',
    embedding: 'bge-m3',
  },

  rag: { profile: 'balanced' },  // 'fast' | 'balanced' | 'precise'

  connectors: [
    { type: 'github', repo: 'org/repo', token: process.env.GITHUB_TOKEN },
    { type: 'notion', token: process.env.NOTION_TOKEN },
  ],

  plugins: [
    '@opendocuments/parser-pdf',
    '@opendocuments/parser-docx',
  ],
})

API keys are stored in .env and loaded automatically.

MCP Server

Use OpenDocuments as a knowledge base for AI coding tools:

{
  "mcpServers": {
    "opendocuments": {
      "command": "opendocuments",
      "args": ["start", "--mcp-only"]
    }
  }
}

19 tools available: search, ask, index, document management, connector sync, stats.

Docker

docker compose up -d                         # Cloud LLM
docker compose --profile with-ollama up -d   # Local LLM

RAG Profiles

| | fast | balanced | precise | |--|------|----------|---------| | Speed | ~1s | ~3s | ~5s+ | | Search depth | 10 docs | 20 docs | 50 docs | | Reranking | Off | On | On | | Cross-lingual | Off | KR + EN | KR + EN | | Hallucination guard | Off | Checks | Strict |

Links

• GitHub: https://github.com/joungminsung/OpenDocuments
• Documentation: https://joungminsung.github.io/OpenDocuments
• Changelog: https://github.com/joungminsung/OpenDocuments/releases
• Issues: https://github.com/joungminsung/OpenDocuments/issues
• Security: https://github.com/joungminsung/OpenDocuments/blob/main/SECURITY.md

License

MIT