Caelune

A cautious adoption judgment is that Caelune looks like a local-first private document retrieval tool, not a general cloud knowledge base, team wiki, or full RAG SaaS product. The current evidence is mostly the official GitHub repository description, which supports the core scope: semantic/vector retrieval for Markdown, PDF, and Tika documents, delivered as a Windows desktop app plus a read-only MCP server. Public evidence is still limited, so confidence should stay moderate.

In practical terms, its role is to add a retrieval layer over your own local files: index scattered documents, make them semantically searchable, and expose them in read-only form to MCP-compatible AI clients. A more accurate analogy is “a local document retrieval engine with an MCP read-only bridge,” not an AI chatbot itself and not a full enterprise knowledge management suite.

On cost and setup, the evidence only supports that it is open source and Windows-oriented; there is not enough evidence to claim stable pricing, hosted service fees, or official API costs.