Haystack 2.0 – compound LLM pipelines are here

I can vouch for Haystack and I'm very happy to see this news. I have started using it to build my company internal RAG platform. It feels way more confortable than using other LLM orchestration looks like LangChain or LllamaIndex, they always felt way too bleeding edge for me with too many breaking changes.

Reading and working with Haystack code was a pleasure and I could even contribute a few time to solve some of my issues or missing features ! A very good foundation for building LLM app. Also it's backed by a company called DeepSet and Haystack exists for multiple years so I guess it's here to stay :)

This represents a major rewrite of the package with more powerful features than ever.

We have introduced the new concept of components. Components are composable and customisable and can be connected into pipelines. Pipelines are dynamic execution graphs that support a range of flows from simple linear chains to more complex execution flows containing loops. This means, you can get started easily with a few lines of code, but have room for extending and customising the logic of the pipeline.

This restructuring of the package paves the way to building truly extensible and composable AI systems ready for production.

The team will be around for questions!

The pipeline framework and flexibility has been a game changer for my hobby projects, its incredibly easy to prototype new LLM logic through the pipeline, swap resources through adjusting document stores, and swap out different models through the generator interface. I've been impressed by the rollout of different integrations within the same interface, and I'm looking forward to seeing where the project goes from here.