Onyx overview and key facts (October 2026)

Onyx is an open-source platform for enterprise search and AI chat that connects to more than 50 business apps and can run fully on a company's own infrastructure. IT teams use it to give employees answers from internal knowledge with any language model.

Onyx indexes content from connected apps and offers search, chat, custom agents with actions, deep research, web search and code execution in a sandbox. It is available on web, desktop, Slack, Discord, as a Chrome extension and as an MCP server. The Community Edition is MIT-licensed; an Enterprise Edition and a hosted cloud version cost extra. The GitHub repository had about 32,000 stars and a push on September 28, 2026.

• Answers from company knowledgeOnyx retrieves relevant passages from indexed apps and lets the language model answer with them. Employees get answers from internal documents instead of generic model knowledge.
• Builds specialized agentsTeams create agents with a defined knowledge subset and actions, for example for support or sales questions. Each team gets an assistant limited to the content it needs.
• Runs deep researchDeep research runs multi-step searches and writes a report with sources. Analysts save time on first overviews of a topic.

• Can run self-hosted and air-gapped with Docker, Kubernetes, Helm or Terraform, so data stays on your infrastructure.
• Model-agnostic: works with commercial and self-hosted language models.

Open-source project that companies run themselves; no lead market is documented.

Self-hosting means your team handles updates, security patches and the search index. Check which features are only in the Enterprise Edition, such as SAML SSO. Connected apps bring their access rights into the index, so test permissions before rollout. Costs for LLM usage come on top.

No reviews yet. Be the first to share how your team uses it.

Every review shows its verification level. We read each one before it goes live. How we check reviews