explaingit

onyx-dot-app/onyx

📈 Trending★ 29,484PythonAudience · developerComplexity · 4/5ActiveSetup · hard

TLDR

Self-hosted AI chat platform that connects to any language model and searches your organization's documents to answer questions accurately.

Mindmap

mindmap
  root((Onyx))
    What it does
      Chat interface
      Document search
      Research flows
      Code execution
    Capabilities
      Voice input/output
      Image generation
      Web search
      Custom agents
    Integration
      50+ data sources
      MCP protocol
      Multiple AI providers
      Local or hosted models
    Team features
      User management
      Single sign-on
      Role-based access
      Usage analytics
    Deployment
      Self-hosted
      Docker support
      Privacy control
      Infrastructure agnostic

Things people build with this

USE CASE 1

Deploy an internal AI assistant that searches your company's documents and knowledge base to answer employee questions.

USE CASE 2

Build a customer support chatbot that uses your product documentation and past tickets to resolve issues automatically.

USE CASE 3

Create a research tool that runs multi-step investigations across web sources and internal data, then generates reports.

USE CASE 4

Set up a code assistant that executes and tests code snippets in a sandbox while accessing your team's codebase.

Tech stack

PythonDockerLLM APIsRAGMCP protocol

Getting it running

Difficulty · hard Time to first run · 1h+

Requires Docker, LLM API keys, document indexing setup, and potentially multiple service components (vector DB, search backend).

License could not be detected automatically. Check the repository's LICENSE file before use.

In plain English

Onyx is an open-source AI chat platform that you can install and run on your own servers. The problem it solves is giving teams a feature-rich interface for interacting with large language models (the AI systems behind tools like ChatGPT) without being locked into any single provider. You connect it to whichever AI model you prefer, whether hosted by a company or running locally on your own hardware, and Onyx provides a polished chat interface on top. Beyond basic chat, it supports RAG (retrieval-augmented generation), which means it can search through your organization's documents and use what it finds to answer questions more accurately. It also supports deep research flows that run multiple steps of investigation before producing a report, web search to get current information, code execution in a sandbox environment, voice input and output, and image generation. You can connect it to over 50 external data sources via built-in connectors or the MCP protocol (a standard for AI-tool integrations). For teams, it includes user management with single sign-on, role-based access control so different people see different resources, usage analytics, and the ability to build custom AI agents with specific knowledge and behaviors. You would use this when your organization wants an internal AI assistant that works with your own data, can be hosted on your own infrastructure for privacy, and works with multiple AI providers. It is written in Python and can be deployed using Docker.

Copy-paste prompts

Prompt 1
How do I set up Onyx to connect to my local Ollama instance and index my company's internal documents?
Prompt 2
Show me how to configure role-based access control in Onyx so different teams only see their relevant data sources.
Prompt 3
What's the process for adding a custom data connector to Onyx if my data source isn't in the built-in 50+ list?
Prompt 4
How do I deploy Onyx on my own servers using Docker and set up single sign-on with my organization's identity provider?
Prompt 5
Can I use Onyx to build a custom AI agent that combines web search, code execution, and document retrieval in one workflow?
Open on GitHub → Explain another repo

Generated 2026-05-18 · Model: sonnet-4-6 · Verify against the repo before relying on details.