deepset-ai/haystack-cookbook

What it solves

Haystack Cookbook addresses the gap between reading documentation and actually building. It provides ready-to-run example notebooks that show how to combine different model providers, vector databases, retrieval techniques, and new experimental features with Haystack. Instead of figuring out integrations from scratch, you can follow these small, focused demos as starting points.

How it works

The project is simply a public repository full of Jupyter notebooks. Each notebook typically showcases a specific small demo or use case, such as using a particular model provider, setting up a vector database, or applying a retrieval technique. The notebooks are meant to be guidelines, not full applications. If you want deeper learning, the README points you to the official Haystack Docs and Tutorials as well as the deepset blog for additional examples.

Who it’s for

This cookbook is for developers who already know they want to build applications with Haystack and need concrete, runnable examples. It suits anyone exploring different providers, vector databases, or retrieval methods and looking for practical templates to adapt. It’s also useful for contributors who want to share their own Haystack examples with the community.

Highlights

• A collection of example notebooks using the Haystack framework
• Covers multiple model providers, vector databases, retrieval techniques, and experimental features
• Most examples are small demos that can be easily adapted
• Contribution process requires opening an issue first, then linking a PR to an assigned issue