Haystack Cookbook
Free👩🏻🍳 A collection of example notebooks using Haystack
FreeFree tier
#Jupyter Notebook
About Haystack Cookbook
Haystack Cookbook is a collection of 47 example notebooks that demonstrate how to use the Haystack framework with various model providers, vector databases, retrieval techniques, and more. Each recipe focuses on a specific capability, such as advanced retrieval, agent building, evaluation, multimodal RAG, data processing, and integration with tools like OpenStreetMap, MongoDB Atlas, Bright Data, Apify, and NVIDIA NIMs. The notebooks are open source and maintained on GitHub, serving as practical guidelines for developers building search and question-answering systems.
Key Features
47 recipes covering diverse topics such as RAG, agents, evaluation, multimodal, and data processing
Hands-on example notebooks with code for immediate experimentation
Integrations with model providers (e.g., OpenAI, LLama, Gemma), vector databases (Chroma, Qdrant, PostgreSQL), and external APIs (Apify, Bright Data, OpenStreetMap)
Filters to explore by topic (Agents, RAG, Evaluation, Multimodal, etc.)
Open source and community-contributed, hosted on GitHub
Pros & Cons
Pros
- Comprehensive range of examples covering many modern NLP tasks and integrations
- All notebooks are open source and freely available
- Easy to filter by topic and sort by newest or last updated
- Notebooks are self-contained demos, making them easy to adapt for your own projects
Cons
- Requires familiarity with the Haystack framework to fully utilize the examples
- Some examples are labeled as experimental and may not be production-ready
- Limited documentation beyond the notebook code itself
Best For
Building retrieval-augmented generation (RAG) pipelines with advanced techniques like query decomposition, HyDE, and metadata enrichmentCreating task-specific agents for itinerary planning, GitHub PR creation, issue resolution, and newsletter sendingEvaluating RAG pipelines using RAGAS, DeepEval, and Prometheus 2Multimodal text generation and audio processing for podcast transcriptionSQL database querying via natural language chat agentsWeb scraping and crawling for question answering
FAQ
What is the Haystack Cookbook?
The Haystack Cookbook is a collection of 47 example notebooks maintained by deepset and the community. It shows how to use the Haystack framework with various model providers, vector databases, retrieval techniques, and more.
How can I contribute to the Cookbook?
The notebooks are maintained in an open source repository on GitHub. You can contribute by submitting pull requests or suggestions via the repository.
What topics are covered in the Cookbook?
Topics include Advanced Retrieval, Agents, Async, Audio, AzureAI Search, Chat Customization, Data Processing, Data Scraping, Evaluation, Function Calling, Guardrails, MCP, Memory, Metadata, Model Serving, Multilingual RAG, Multimodal, Observability, OpenStreetMap, Prompting, RAG, SQL, Summarization, Vector Databases, and Web-QA.