👩🏻🍳 Haystack Cookbook
🧑🍳🍳Discover The Haystack Cookbook here
A collection of example notebooks using Haystack 👇
You can use these examples as guidelines on how to make use of different model providers, vector databases, retrieval techniques and more with Haystack. Most of them showcase a specific, small demo.
To learn more about how to use Haystack, please visit our Docs and official Tutorials.
For more examples, you may also find our Blog useful.
Note: Unless '(Haystack 1.x)' is mentioned in the title, all of these examples use Haystack 2.0 onwards.
| Name | Colab |
|---|---|
| Improving Retrieval with Auto-Merging |  |
| Speaker Diarization with AssemblyAI | |
| Advance Prompt Customization for Anthropic | |
| Advanced RAG: Query Decomposition and Reasoning | |
| Advanced RAG: Automated Structured Metadata Enrichment | |
| Techcrunch News Digest with Local LLMs using TitanML Takeoff | |
| Use Gemini Models with Vertex AI | |
| Gradient AI Embedders and Generators for RAG | |
| Mixtral 8x7B with Hugging Face TGI for Web QA | |
| Amazon Bedrock and OpenSearch for PDF QA | |
| Use Zephyr 7B Beta with Hugging Face for RAG | |
| Hacker News RAG with Custom Component | |
| Use Chroma for RAG and Indexing | |
| Using the Jina-embeddings-v2-base-en model in a Haystack RAG pipeline for legal document analsysis | |
| Multilingual RAG from a podcast with Whisper, Qdrant and Mistral | |
| Improve retrieval by embedding meaningful metadata | |
| Advanced RAG: Query Expansion | |
| Information extraction via LLMs (Gorilla OpenFunctions) | |
| Information extraction via LLMs (NexusRaven) | |
| Using AstraDB as a data store in your Haystack pipelines | |
| Streaming model explorer: compare how different models handle the same prompt. | |
| Function Calling with OpenAIChatGenerator | |
| Use the vLLM inference engine in Haystack 2.x | |
| Build with Google Gemma: chat and RAG | |
| Optimizing Retrieval with HyDE | |
| RAG pipeline using FastEmbed for embeddings generation | |
| Sparse Embedding Retrieval with Qdrant and FastEmbed | |
| Hybrid Retrieval: BM42 + Dense Retrieval (with Qdrant and FastEmbed) | |
| Air-Gapped RAG pipelines with NVIDIA NIMs | |
| Evaluate a RAG pipeline using Haystack-UpTrain integration | |
| RAG on the Oscars using Llama 3.1 models | |
| Chatting with SQL Databases | |
| Evaluate a RAG pipeline using DeepEval integration | |
| Evaluate a RAG pipeline using Ragas integration | |
| Extract Metadata Filters from a Query | |
| Run tasks concurrently within a custom component | |
| Prompt Optimization with DSPy | |
| RAG Evaluation with Prometheus 2 | |
| Build quizzes and adventures with Character Codex and llamafile | |
Invoking APIs with OpenAPITool |
|
| Extract and use website content for RAG with Apify | |
| Analyze Your Instagram Comments’ Vibe with Apify and Haystack | |
| Conversational RAG using Memory | |
| Evaluating RAG Pipelines with EvaluationHarness | |
| Define & Run Tools | |
| Agentic RAG with Llama 3.2 3B | |
| Create a Swarm of Agents | |
| Cohere for Multilingual QA (Haystack 1.x) | |
| GPT-4 and Weaviate for Custom Documentation QA (Haystack 1.x) | |
| Whisper Transcriber and Weaviate for YouTube video QA (Haystack 1.x) | |
How to Contribute a Cookbook
If you have an example that uses Haystack, you can add it to this repository by creating a PR. You can also create a PR from Colab by creating a Fork of this repository and selecting "Save a Copy to GitHub". Once you add your example to your fork, you can create a PR onto this repository.
- Add your Notebook
- Give a descriptive name to your file that includes the names of (if applicable) the model providers, databases the technologies you use in your example and/or the task you are completing in the example.
- Make sure you add it to
index.tomlincluding its title and topics. - Make sure to add a row in the table above 🎉