Haystack AI Framework

⭐ Haystack AI Framework just crossed 25,000 stars on GitHub! This number means a lot to us, but what it really represents is you. Every contributor who opened a pull request. Every community member who answered a question on Discord. Every developer who filed an issue, wrote a notebook, gave a talk, or simply built something incredible with Haystack and shared it with the world. It's a true community effort 💙 When we first started Haystack, we believed that the journey to building great AI applications should be open, composable, and community-driven. 25,000 stars later, that belief has never felt more validated. Thank you to every single one of you: contributors, users, advocates, and builders. Let’s keep going! 🚀 → Join our stargazers on GitHub: https://lnkd.in/dn-BF-e #haystack #opensource #python deepset, makers of Haystack

Haystack pipelines need strong retrieval primitives: embedders, rerankers, extractors. Today we're announcing the native Haystack and SIE integration from Superlinked: drop-in Haystack components for every stage of the pipeline, self-hosted on 85+ open-source models. 🐍 pip install sie-haystack 🔗 Documentation: https://lnkd.in/dCjMFZ9J

deepset is heading to London with NVIDIA 🇬🇧 On April 29th, we're co-hosting part 3 of our Live Unconference Series at the Balderton Offices. We’ll dig into one of the most important challenges in AI right now: context engineering - the practice of structuring and delivering the right information to a model at the right time, which often determines whether AI systems actually work in production. We are excited to explore this together with the London AI engineering community, powered by Haystack AI Framework and NVIDIA Nemotron models, including a live demo on the night. No slides, no long talks. The Unconference format means everyone comes as a peer, to share what's worked, and learn from the real experiences of others in the room. Topics are shaped entirely by what attendees bring to the table. Whether you're already in the weeds or just getting started with AI challenges, come and see what the London AI engineering scene is building. 📅 April 29th, 18:00 - 20:30 📍 Balderton Offices, London Register now using the link in the comments.

⚡ Thunderbolt is trending on GitHub today - and we're proud to be Mozilla's launch partner. MZLA (Mozilla's subsidiary behind Thunderbird) just launched Thunderbolt: an open-source, self-hostable AI client for enterprises and public sector organizations that need sovereignty over their AI stack. Native apps across web, desktop, and mobile. Together, Thunderbolt + Haystack Platform deliver a complete sovereign AI stack. Thunderbolt and Haystack are both open-source, and our community proved you can use both tools already. Integration docs are in the first comment down below. 👇

Haystack 2.28 is here 🚀 This release makes agent tool development more flexible and document processing more precise. Many interesting updates this time, but only one highlight: 🔗 Pass Agent State Directly to Tools No more manual wiring. Tools and components can now declare a state parameter and receive the live agent State at runtime automatically — giving them full access to conversation history and context without extra connections. 💙 Big thanks to our contributors to this release: Amanpreet Singh, Dina Deifallah , PhD, Max Swain, Nhat Huy Vu, Sriniketh Jayasendil, and others! 👇 Full release notes in the comments

Local Gemma 4 agent: drop in a map, get the location, live weather, and top spots to visit I've been exploring what Gemma 4 can do in a local agentic setup and put together a 📓 𝙣𝙤𝙩𝙚𝙗𝙤𝙤𝙠 with Gemma + Haystack AI Framework covering 4 demos. Another interesting one is the 𝗚𝗶𝘁𝗛𝘂𝗯 𝗔𝗴𝗲𝗻𝘁. I initially tried to load all tools from the GitHub MCP server, quickly filling the context available on Colab -> unusable, forgetful agent ❌ Then I used the 𝗦𝗲𝗮𝗿𝗰𝗵𝗮𝗯𝗹𝗲 𝗧𝗼𝗼𝗹𝘀𝗲𝘁 🔎 🧰  It dynamically discovers the right tools from the GitHub MCP server on the fly, loading only what it actually needs for the task at hand, keeping context lean. Now it actually works. (Thanks Vladimir Blagojevic for this abstraction) The notebook also contains 💎 Multimodal weather agent: the mystery map demo above 💎 Visual Question Answering from a paper 💎 RAG on Rock music 📓 notebook below (Google AI for Developers)

Bigger context windows don't mean better agents - they mean more competition for the model's attention. New post: context engineering for agentic systems by Kacper Łukawski. What fills the context window, why bloat hurts quality and cost, and how Haystack gives you full control over your agent harness - the infrastructure layer that keeps it all under control. 🔗 https://lnkd.in/dJSCUW3e Part 1 of a series - what would you most want us to cover next?

New tutorial: compress your LLM's KV cache with TurboQuant + Haystack 🗜️ by Kacper Łukawski We used turboquant-vllm (a community TurboQuant implementation, based on the Google Research recent paper) hooked into Hugging Face chat generator. Since Haystack is provider-agnostic, anything compatible with Hugging Face transformers, Ollama, or vLLM works just as well. 👉 https://lnkd.in/dm7FC4fb

⚡ We're proud to partner with Mozilla on Thunderbolt to deliver a complete sovereign AI stack for enterprises and public sector organizations that need full control over how AI is built, deployed, and used. Together, Thunderbolt and Haystack AI Framework give organizations a self-hostable AI client paired with the infrastructure to run RAG pipelines, agents, and AI workflows on your organization's actual data. Deployed within your infrastructure, free from vendor lock-in. Learn more about Thunderbolt and our partnership here: https://lnkd.in/g_-uQDwg

MarkItDown by Microsoft quickly became one of the most talked-about document conversion libraries in the Python ecosystem. The reason is simple: it just works. PDF, DOCX, PPTX, XLSX, HTML, images - all converted to clean Markdown, locally, with no external API calls. For RAG pipelines, that matters more than people realize. The quality of your conversions directly shapes the quality of your retrieval. Use MarkItDownConverter to convert virtually any file format into Haystack Document objects with Markdown content - standalone or wired directly into your indexing pipelines. Everything runs locally, so your data stays yours. 🐍 pip install markitdown-haystack 🔗 Documentation: https://lnkd.in/gJfaVi8H