Haystack
A Python framework for assembling document indexing, retrieval, reranking, and answer pipelines.
Published by deepset
How to use it
Build an indexing pipeline that prepares and writes documents, then a separate query pipeline that retrieves passages. This keeps document updates independent of requests for answers.
Use named inputs and outputs to inspect the results of individual components. A pipeline can run with local components when its embedding model, store, and generator support local execution.
- Compatibility
- Python, Local inference, Hosted integrations
- Components
- RAG pipeline, Document indexing, Retrieval orchestration
- Access
- Local Python package. Hosted model and search integrations can require credentials.