Haystack

Open-source AI orchestration framework by @deepset_ai.
Build context-engineered agents & RAG systems in Python.

Discord for support →

[Haystack @Haystack_AI]

One Name. One Product Family. One Look 💙

We’re unifying the Haystack ecosystem at @deepset_ai under one name and a new logo, reflecting its role as a framework, a community, and the foundation of our enterprise platform.

👉 Read the announcement: https://t.co/AmA2jhnXDE

[Haystack @Haystack_AI]

🧠 @cognee_ is now available as a memory store integration in Haystack. Seed permanent knowledge into your graph and then retrieve relevant memories in agentic pipelines. With Cognee and Haystack, you can:

- Write domain knowledge once to the permanent knowledge graph
- Build agents that combine conversation history with long-term facts
- Create session-scoped stores with just a session_id string

This makes it simple to build stateful agents that remember facts across sessions and grow smarter as they interact, combining real-time conversation with persistent domain knowledge in a single unified system.

🐍 pip install cognee-haystack
🔗 https://t.co/wHaS3n048A

[Haystack @Haystack_AI]

Another integration alert! Use FunASRTranscriber to convert audio files to Documents with fully self-hosted speech recognition supporting 50+ languages. No API keys required.

FunASR handles transcription, speaker diarization, and timestamp extraction. Pair it with Haystack pipelines to build audio-to-document workflows that stay entirely on your infrastructure. Models are downloaded on first use and cached locally.

@ModelScope2022

🐍 pip install funasr-haystack
🔗 https://t.co/JwR1930PK3

[OrcaRouter 🐳 @OrcaRouter]

Our integration just got merged into @Haystack_AI , try it out!

[Haystack @Haystack_AI]

ArangoDB is available as a Document Store in Haystack. Use ArangoDocumentStore to connect directly to your ArangoDB instance, and integrate it into RAG pipelines that leverage ArangoDB's multi-model capabilities.

@arangoai flexible schema and native support for documents, graphs, and vectors make it ideal for applications that need both structured data storage and semantic search. Whether you're indexing documents, managing relational data with embeddings, or building hybrid search experiences, ArangoDB handles it seamlessly.

🐍 pip install arangodb-haystack
🔗 https://t.co/2kDTPiga09

[Bilge @bilgeycl]

↔️ MCP works in both directions with @Haystack_AI

Your Haystack agent can consume tools from external MCP servers. And your Haystack pipelines/agents can be exposed as MCP tools for any MCP-compatible assistant.

I wrote a guide covering three approaches:

🔹 Haystack as an MCP client — use MCPTool, MCPToolset, or SearchableToolset to give your agent access to external tools

🔹 Haystack as an MCP server with Hayhooks — self-host your pipelines and agents and expose them to Claude, ChatGPT, Cursor, or any MCP-compatible client

🔹 Haystack as a managed MCP server on Haystack Enterprise Platform — production-grade MCP support with access control and observability built in. This is also how we built the Haystack Docs MCP Server at @deepset_ai

Read the full article here 👇
https://t.co/pct266NF2T

[Haystack @Haystack_AI]

A new Document Store has been added to Haystack. Use @supabase to build pgvector vector search capabilities directly into your Haystack pipelines, use Groonga for efficient full-text search, or choose Supabase Storage for seamless file handling.

@supabase makes it easy to host your vector database in the cloud without managing infrastructure - just spin up a project, get connection credentials, and you're ready to embed and search. This works well for applications needing scalable vector storage with the familiarity and power of @PostgreSQL

🐍 pip install supabase-haystack
🔗 https://t.co/3nOgSxyYNe

[Haystack @Haystack_AI]

Extract text, tables, and forms from images and PDFs with Amazon Textract in Haystack. AmazonTextractConverter brings @awscloud's OCR capabilities directly into your document processing pipelines. Just pass an image or single-page PDF and ask natural-language questions to extract structured data automatically.

🐍 pip install amazon-textract-haystack 🔗 https://t.co/2M8wk3gsNu

[Haystack @Haystack_AI]

👋 @LiteLLM is now available as a Chat Generator in Haystack. Access 100+ LLM providers from a single interface - OpenAI, Anthropic, Groq, Cohere, and more. Switch models and providers without rewriting code.

🐍 pip install litellm-haystack
🔗 https://t.co/yUM7oZMo91

[Haystack @Haystack_AI]

@m13v_ this hasn't been an issue in our tests, but feel free to open an issue in the Haystack repo with more details
https://t.co/dhoTPHTRiK
we're happy to improve the implementation 🙌

[Haystack @Haystack_AI]

Haystack now publishes a public Model Context Protocol server. Point Claude Code, Cursor, or any MCP-compatible coding agent at it and get real-time access to the latest Haystack docs - no API keys or sign-ups needed.

🖇️ https://t.co/hb59y1733l

[Haystack @Haystack_AI]

@rryssf MCP server returns the most relevant 5 chunks (chunks are by H1/H2 level), not the whole content, unless the agent decides to fetch the page content through the link attached to the chunks

Give it a try and let us know if you have any issues, happy to improve the implementation!

[Haystack @Haystack_AI]

@m13v_ Doc pages are chunked by H1/H2 level at index time & every query reranks to the top 5 chunks via a Haystack pipeline running in the backend.

URLs for the full docs page are included in the chunks, so if an agent needs the entire page, they can fetch it directly from the link.

[Haystack @Haystack_AI]

💙 Big thanks to our contributors to this release!

👇 Full release notes https://t.co/hF9UUzphsl

[Haystack @Haystack_AI]

Haystack 2.30 is here 📷

Smarter code processing and simpler APIs. Two highlights this time.

[Haystack @Haystack_AI]

💬 Simpler ChatGenerator API

All Haystack ChatGenerator components now accept a plain string for the messages parameter, automatically wrapping it in a ChatMessage with the user role. This one-line change eliminates boilerplate and makes switching from Generator to ChatGenerator seamless - no more manual ChatMessage construction. Applies to AzureOpenAIChatGenerator, HuggingFaceAPIChatGenerator, OpenAIChatGenerator, and others, with more rolling out in Haystack Core Integrations soon.

[Haystack @Haystack_AI]

Vespa is now available as a Document Store in Haystack. Use VespaDocumentStore for hybrid and semantic search with a powerful, production-ready engine, and pair it with VespaEmbeddingRetriever to index and retrieve documents directly in your pipelines. Metadata filtering included.

@vespaengine excels at large-scale information retrieval with advanced features like real-time indexing, multi-modal search, and distributed document management - ideal for applications that demand both speed and sophistication.

🐍 pip install vespa-haystack
🔗 https://t.co/Z1pz3A32tz

[Haystack @Haystack_AI]

Applications are open now. We review quarterly, so the sooner you apply the better! ✅

You've been showing up. We want to show up for you.

https://t.co/dDUoyky4yU

[Haystack @Haystack_AI]

We're launching the Haystack Ambassador Program 🎉

It's for the people already building, teaching, contributing to the Haystack ecosystem and who want to go further.

🧵 Here's what it is, who it's for, and how to apply

[Haystack @Haystack_AI]

What you get 🫱

→ Direct access to the Haystack core team
→ Early previews and feedback on what you're building
→ Support for talks, workshops, and meetups
→ Travel support when community growth is on the line
→ Amplification of your content through our channels