haystack
https://github.com/deepset-ai/haystack
Python
:mag: LLM orchestration framework to build customizable, production-ready LLM applications. Connect components (models, vector DBs, file converters) to pipelines or agents that can interact with your data. With advanced retrieval methods, it's best suited
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported5 Subscribers
Add a CodeTriage badge to haystack
Help out
- Issues
- OpenAIResponsesChatGenerator: streamed reasoning and ChatMessage.reasoning join summary parts differently
- fix: handle chat generator replies without text or with no replies in LLMEvaluator and LLMMetadataExtractor
- DocumentSplitter: Hindi, Marathi, Bengali and other danda (।) scripts aren't split into sentences, so the whole document becomes one chunk
- test: relax pdf content extraction e2e test
- Function tools pass raw LLM JSON to the function: Pydantic/dataclass parameters arrive as dicts and Literal/int arguments are never checked
- fix: serialize pydantic models in JSON mode so snapshots with datetime fields can be saved
- fix: split sentences ending with danda (।, ॥) in DocumentSplitter
- fix: accept signed offsets in the Jinja2 {% now %} tag
- Pipeline.connect rejects a Literal output connected to an input of its value type (e.g. Literal["a", "b"] -> str)
- docs: add docs for Amazon S3 Vectors integration
- Docs
- Python not yet supported