AutoRAG
View on GitHubAutoRAG: Now your agent can find anything in your computer. It gets smarter if you are using it frequently.
A TypeScript librarian agent that searches local files and federated data sources using BM25, vector, and hybrid retrieval, then reads source documents to verify and cite answers. Includes a model-free CLI retrieval mode.
Use Cases
Search local documents and answer questions with source citationsRetrieve evidence across chat archives, email, Slack, and NotionRun local model-free BM25 or vector retrieval for agent workflowsCurate findings and improve retrieval choices using feedback
Built With
- Language
- TypeScript
- Frameworks
- Pi
Tags
librarian-agent · document-search · RAG · local-first · semantic-search · BM25 · hybrid-search · embeddings · knowledge-curation · datasource-federation · agent-skills · CLI