onyx-dot-app/onyx is an open-source AI platform that indexes information from connected applications to provide the knowledge that teams and AI agents need. The latest release left in the repository record is v4.9.2. However, the provided material contains neither release notes nor changes for this version, so it cannot be confirmed what went into it. The sources for the version and the features are the onyx-dot-app/onyx repository on GitHub and its README.
What Onyx does is index organizational documents so they can be searched, and it provides answers with attached evidence. The README states that it supports hybrid-index-based Agentic RAG (a method of answering based on documents found through search), Deep Research, custom agents, web search, external actions, and MCP integration. A self-hosted deployment that performs document indexing, the database, and data processing in your own environment is also possible. Because there is an interface for browsing search results and evidence citations, you can set up a flow of indexing documents and then receiving answers with evidence attached. The repository description says it is an AI chat platform that works with all LLMs. The implementation language is Python, the repository was created on April 27, 2023, and it has 32610 stars.
That said, the verifiable scope is narrow. Since the material has no change history for v4.9.2, there is no way to know what this release improved. The repository license field is NOASSERTION, and the MIT license applies to the Onyx Community Edition named in the README. No benchmark comparing search quality or response performance is presented either. In the evaluation record, only a momentum score of 5.4 out of 10 remains, and the freshness, technical and quality items are empty. So if you are evaluating Onyx as an in-house document search tool, it makes sense to check the original license text and the v4.9.2 release notes first.