Onyx
Open-source AI platform with advanced RAG, connectors for 50+ data sources, and agents
4.4/ 530,859 GitHub starsPython/TypeScriptMIT
⚡ TL;DR
What
Open-source AI platform with advanced RAG, connectors for 50+ data sources, and agents
Who
Enterprises wanting AI across all data sources, Teams needing RAG with 50+ connectors, Companies with data sovereignty requirements
Catch
Resource-heavy
Verdict
⭐⭐⭐⭐ Essential
🎯 The Problem It Solves
You want a self-hosted AI chat that connects to all your company's data sources (Slack, Google Drive, Confluence, GitHub, etc.) with enterprise-grade RAG and agent capabilities. Onyx gives you a complete AI platform with 50+ built-in connectors.
🔧 How It Works
Onyx is a Python/TypeScript web app that indexes data from 50+ enterprise sources (Slack, Google Drive, Confluence, GitHub, Jira, etc.) and provides a chat interface with advanced RAG. Agents can search across all connected sources, answer questions with citations, and perform multi-step reasoning.
🚀 Installation & Quick Start
Installation
docker run -d -p 3000:3000 onyx/onyx:latestQuick Start
- Deploy with Docker
- Connect your data sources (Slack, Drive, GitHub, etc.)
- Ask questions across all your data
- Create agents for automated tasks
✅ Pros
- •50+ data connectors out of the box
- •Enterprise-grade RAG
- •Agent capabilities
- •Self-hosted option
- •Active development
❌ Cons
- •Resource-heavy
- •Complex setup
- •Some features require enterprise version
💬 Practitioner Verdict
"Onyx is the most complete self-hosted AI platform for enterprises. The 50+ connectors and advanced RAG make it the go-to choice for companies wanting AI across all their data."
— AI RepoIndex, Security Reviewer
📊 Specifications
- Language
- Python/TypeScript
- License
- MIT
- Platform
- Linux, macOS, Windows
- Kill Chain
- None
💰 Pricing Reality
Open-source core (MIT). Enterprise version available. Self-hosted or cloud.
👥 Community Health
Stars30,859
Forks3,857
Contributors617
Health Score8.5/10
🏷️ Tags
Open SourceSelf-HostedRAG