onyx-dot-app/Onyx
Open Source AI Platform - AI Chat with advanced features that works with every LLM
Onyx is an open-source AI chat and enterprise-search platform that connects to your company's documents and apps, then answers questions over them with any LLM. Self-hosted by design, it gives teams a private, citable knowledge assistant without sending data to a third-party cloud.
Our Review
Onyx (formerly Danswer) has gathered 29,000 GitHub stars as a self-hosted alternative to closed enterprise-search and AI-chat products. It connects to your company's apps and documents, indexes them, and lets employees ask questions in chat with citations, all running on infrastructure you control.
What Onyx does:
- •Connects to your sources 40+ connectors pull in docs from Google Drive, Slack, Confluence, Notion, GitHub, and more, kept in sync.
- •AI chat over your data ask questions in a ChatGPT-style UI and get answers with citations back to the source document.
- •Enterprise search a unified search layer across all connected apps, with permission-aware results.
- •Works with any LLM plug in OpenAI, Anthropic, or local models via Ollama and swap providers without lock-in.
- •Custom assistants build purpose-specific assistants with their own prompts, sources, and tools for different teams.
- •Self-hosted and permission-aware runs in your environment and respects each user's document access rights.
Getting started:
Self-host with Docker Compose from the repo, or use Onyx Cloud. Add a connector, point it at your data, pick a model, and start chatting. Docs at docs.onyx.app.
Limitations:
Onyx uses a mixed license: the core is open source (MIT), but some enterprise features sit under a separate commercial license, so a self-hosted deployment may not match Onyx Cloud or the enterprise edition. Running it means operating the stack (it uses Postgres and a vector store) yourself. Indexing many large sources takes compute and tuning, and the breadth of connectors and admin features is more than a single-user setup needs.
Our Verdict
Onyx is one of the strongest open-source options in 2026 for a private, company-wide AI assistant. If you want ChatGPT-style answers grounded in your own Slack, Drive, and Confluence, with citations and without shipping data to a vendor, Onyx delivers that self-hosted, with 29,000 stars behind it.
For developers and IT teams, Onyx self-hosts with Docker, connects to 40+ data sources, and works with any LLM including local models, so you control both the data and the model. Permission-aware search means results respect each user's existing access, which matters for real enterprise rollouts.
Skip Onyx if you only need a single-user chatbot or have no internal documents to search; a hosted assistant is far less to run. If you want a lighter self-hosted chat UI without the enterprise-search and connector layer, Open WebUI is simpler to operate.
Frequently Asked Questions
What is Onyx?
Onyx, formerly known as Danswer, is an open-source AI platform for chat and enterprise search over your own data. It connects to apps like Google Drive, Slack, and Confluence, indexes their content, and lets people ask questions in a chat interface with citations back to the source. It is self-hosted and works with any LLM.
Is Onyx free and open source?
Yes, with nuance. Onyx's core is open source under the MIT license and free to self-host as of 2026. Some enterprise features are under a separate commercial license, and Onyx also offers a hosted Cloud plan. For many teams, the free self-hosted core covers connectors, AI chat, and search.
What data sources does Onyx connect to?
Onyx ships more than 40 connectors for common workplace tools, including Google Drive, Slack, Confluence, Notion, GitHub, Jira, and email, and it keeps them synced. Once connected, their content is indexed and searchable, and answers cite the originating document. Permissions from the source are respected, so users only see what they are allowed to.
How is Onyx different from Open WebUI?
Both are self-hosted, open-source AI chat interfaces. Open WebUI focuses on chatting with models, especially local ones via Ollama. Onyx adds enterprise search: connectors to company apps, indexing, citations, and permission-aware retrieval. Choose Open WebUI for a lightweight local-model chat UI; choose Onyx when you need answers grounded in your organization's documents.
Can Onyx use local LLMs?
Yes. Onyx is model-agnostic as of 2026 and works with any LLM, including hosted providers like OpenAI and Anthropic and local models served through Ollama or compatible endpoints. That lets you run the whole stack privately, keeping both your documents and the model inside your own environment for sensitive use cases.
How do I install Onyx?
Visit the GitHub repository at https://github.com/onyx-dot-app/onyx for installation instructions.
What license does Onyx use?
Onyx uses the MIT license.
What are alternatives to Onyx?
Explore related tools and alternatives on My AI Guide.
Open source & community-verified
MIT licensed: free to use in any project, no strings attached. 30,006 developers have starred this, meaning the community has reviewed and trusted it.
Reviewed by My AI Guide for relevance, quality, and active maintenance before listing.
Topics