deepset, makers of Haystack

Sovereign AI starts with Haystack. The platform to build, run, & govern AI agents on your own terms. Follow @haystack_ai for open-source and community updates.

[deepset, makers of Haystack @deepset_ai]

What a moment for the deepset team at #NVIDIAGTC Berlin! 🚀

We welcomed @JensenHuang to our booth to discuss our work with @nvidia on open, sovereign AI agents and how Haystack is helping scale AI across the @EU_Commission.

And to top it off, Jensen signed our DGX Spark! That one's going straight into the deepset history books.

[deepset, makers of Haystack @deepset_ai]

Space is limited, register here: https://t.co/3VnnrmUGHF

[deepset, makers of Haystack @deepset_ai]

🇨🇦 @deepset_ai and @mozilla are hosting an unconference in Toronto on Nov 17.

As more organizations move AI agents into production, a key question keeps coming up: how do you build AI systems while keeping control of your data, models, and infrastructure?

That’s the focus of 𝘜𝘯𝘤𝘰𝘯𝘧𝘦𝘳𝘦𝘯𝘤𝘦: 𝘖𝘸𝘯 𝘠𝘰𝘶𝘳 𝘈𝘐 𝘚𝘵𝘢𝘤𝘬, an evening at Mozilla’s Toronto office that brings together Thunderbolt ⚡️ and @Haystack_AI

The evening opens with a live demo of the two working together, followed by small-group discussions on topics attendees choose during registration.

Organized in partnership with TorontoAI and the @AgenticAIFdn, join us!

📍 Mozilla Toronto
🗓️ Nov 17, 6PM

[Bilge @bilgeycl]

toronto, let’s talk about your AI stack 🫵 🇨🇦

@deepset_ai is coming to town and together with our friends at @mozilla, we’re hosting Unconference: Own Your AI Stack, with TorontoAI as our community partner 🫶

we’ll run one unconference round with about three sessions, and you will help us shape the topics when you register!

we’ll also kick things off with a live demo of @Haystack_AI working with Thunderbolt ⚡️, showing what an open AI setup can look like in practice

📍 Mozilla Toronto
🗓️ nov 17, 6PM

can't wait to meet you there 👋

[Bilge @bilgeycl]

https://t.co/nyD5YroGcH

[deepset, makers of Haystack @deepset_ai]

The best AI agent delivers the quality you need at a cost you control.

We're at @NVIDIAGTC Berlin demonstrating how Haystack Enterprise Platform helps enterprises build and run production-ready sovereign AI agents.

📍 Find us at booth 3071 in the AI Neighbourhood at STATION Berlin on Wednesday and Thursday.

🎤 On Thursday we’re on stage twice with @nvidia:
• 10:00 Secure, self-improving agents
• 14:00 Trustworthy AI platforms

Come meet us!

#NVIDIAGTC #gtcberlin

[Bilge @bilgeycl]

Engineers building production-scale AI applications, this one's for you 🫵

Next Wednesday (Oct 21), we're opening the Haystack Enterprise Platform to a small group of engineers for one night at the @deepset_ai office in Berlin.

You get:
🔑 Your own platform workspace and API keys
🧭 A walkthrough from the people who build the platform
🛠️ Two hours to build and deploy your own @Haystack_AI agent
🏆 A leaderboard, with a prize for the top score
⏳ Access that stays open for a week, so you can keep going on your own use cases

If you work in public sector or at a large organization, you know how hard it is to move AI agents from prototype to production with the control, security, and governance your organization requires. This is your chance to try a platform built for exactly that, hands-on, with our platform team in the room to help.

Spots are very limited, and we're reviewing every application to build the right room. Bring a laptop; we'll handle the rest.

👉 Register here: https://t.co/T3Um65diR3

[deepset, makers of Haystack @deepset_ai]

And they're available on the Haystack Enterprise Platform!

[Haystack @Haystack_AI]

Most of what we ask LLMs to do inside pipelines isn't generation. It's decisions. @typesafeai System One models are built for exactly that, and they're now in @Haystack_AI

The new integration gives you two components:

🏷️ TypeSafeDocumentClassifier answers your questions about each document and stores the results in its metadata.
🔀 TypeSafeTextRouter sends text to the right branch, with an optional low_confidence output so uncertain cases can fall back to an LLM or a human.

Use them with TypeSafe's hosted Jev model, or keep everything on your own infrastructure with Ollaya, which serves open decision models like Laya locally through the TypeSafe API.

🐍 pip install typesafe-haystack
👉 https://t.co/056SzuQHig

[Haystack @Haystack_AI]

🌐 Building an AI agent that needs web intelligence?

@youdotcom brings semantic web search and real-time retrieval directly into your Haystack agents!

Add YoucomWebSearch to your agent (or your pipeline), configure your https://t.co/lz6MKhEPR4 API credentials, and your agent can search the live web mid-run, pulling in content, snippets, and metadata that provide enough ground for downstream tasks.

Think of an agent that researches a topic in real time instead of working off its index.

🐍 pip install youcom-haystack
🔗 https://t.co/85hmbAj4Px

[Haystack @Haystack_AI]

📣 We're hosting our next unconference in Berlin with @prior_labs, exploring open-weight models and open-source agents! Come join us if you'd like to discuss:

→ Structured data in agentic workflows
→ When agents should hand off to TabPFN for prediction
→ Benchmarking tabular models vs. LLMs
→ Why orchestration matters for production agents
→ What owning your stack end-to-end actually requires
→ Open-weight vs. closed models in practice

Or anything else that's been challenging you lately!

📍 @deepset_ai HQ, Berlin
🗓️ September 24th, 6PM-9PM

Register: https://t.co/dyPd5xvZRr

[deepset, makers of Haystack @deepset_ai]

We’re proud to be named one of the @Siftedeu 100 Fastest Growing Startups in DACH & CEE.

For us, this recognition marks another step in our mission to help organisations build and scale sovereign AI on their own terms.

With Haystack, we’re helping enterprises and public sector teams turn AI systems and agents into production applications that are controllable, adaptable, and built to deliver real business outcomes.

Check out more in the full report: https://t.co/8fwENVtTJj

#Sifted100 #SiftedLeaderboards

[Bilge @bilgeycl]

🗽In NYC tonight? @DylanCouzon & @qdrant_engine are hosting a debate night, and @deepset_ai is teaming up

Hear some hot AI takes, get a chance to win prizes, and learn about @Haystack_AI and @qdrant_engine along the way!

https://t.co/FqJTAyZVGv

[deepset, makers of Haystack @deepset_ai]

We are thrilled to be among the ten finalists for The Spark – The Technology Award.

For eleven years, @handelsblatt and @McKinsey have been recognizing Germany’s most innovative technology startups.

This year’s theme is "Scaling Deep Tech", focusing on technologies that are already making a real-world impact today and shaping the future of industry and the economy.

Now, we need your support.

Vote by September 7th to help us reach the Top 3! Your vote counts directly toward the jury's decision.

Vote for us here: https://t.co/2H3vnBf7oL

[deepset, makers of Haystack @deepset_ai]

Work on a production-grade AI agent is never really finished. You ship it, watch how it behaves, and keep improving it.

But that cycle only works if you can see what is happening inside every run.

Getting that visibility often meant sending prompts, responses, and trace data to a separate third-party tool.

Until now - we are launching Traces in Haystack Enterprise Platform: full observability for every pipeline, built directly into the platform that runs it.

→ See every run end to end: spans, tokens, inputs, outputs, logs, errors
→ Read the root cause in seconds, and tell a config error from a provider timeout
→ Compare two pipeline versions to see what a change cost before you ship it
→ All on your own infrastructure. Your data never leaves.

Build, run, and improve your agents with full visibility at every step - all on one platform.

Read the blog to see Traces in action: https://t.co/oqEce6CVPH

[deepset, makers of Haystack @deepset_ai]

The new “The Top 10 Trends In Sovereign AI, 2026” research authored by @forrester Principal Analyst Dario Maisto explores how organizations are redefining Sovereign AI not as isolation, but as architectural choice.

We're proud to see deepset and the Haystack Enterprise Platform recognized, including our work powering the @EU_Commission's AI@EC Platform.

Our take: Sovereign AI means the freedom to choose your models, infrastructure, and deployment environment, while retaining full control.

Learn more: https://t.co/ndvViHNpAV

#SovereignAI #Haystack #EnterpriseAI

[Haystack @Haystack_AI]

Haystack 3.0 shipped this month. We spent launch week showing you what it can do. Now here's what it took to build it 👇

Since @Haystack_AI 2.0: 2,400+ pull requests merged, 195 contributors, ~170 of them making their first-ever contribution along the way. 17,258 CI runs since the v3 branch was cut, roughly one every five minutes, around the clock, for eight weeks.

Most-edited file of the whole cycle: https://t.co/3LBMRsRLQZ. Of course it was.

Migrating to 3.0? The full guide covers all breaking changes: https://t.co/I7lqvDc0h6
Or skip straight to haystack-v2-to-v3, an agent skill that migrates your v2 pipelines and agents for you.

💙 Huge thanks to our open-source team at @deepset_ai, and to our community contributors who shipped fixes and features for v3 alongside us! Haystack 3.0 is yours as much as it's ours.

Got questions about migrating, or curious why we made a specific call? Join us at Haystack 3.0 Office Hours next week, August 4, with @LukawskiKacper and @bilgeycl
👉🏼 Register: https://t.co/ex2pXiPufK

[deepset, makers of Haystack @deepset_ai]

AI sovereignty has become a mainstream enterprise priority. But the conversation often centers on a single question: Where does the model run? That's only part of the story.

When you deploy AI agents in production, the real value accumulates around the model: the skills that encode how your team works, along with the tools, memory, workflows, and context unique to your organization. Models can be replaced. This agent orchestration layer cannot.

The more important question becomes: Who controls that layer? If your skills, logic, and context only work inside a single vendor's runtime, lock-in hasn't disappeared. It has simply moved from the model to the orchestration layer around it.

Today we're excited to introduce Haystack 3.0, designed to keep that layer open and under your control:
→ Portable skills you can read, version, and own
→ A programmable agent loop you can inspect, audit, and adapt
→ Built-in observability into how your agents think and operate

Every AI agent runs inside an orchestration layer. True AI sovereignty means owning the agent orchestration layer, not just choosing where the model runs.

With 3.0, that control stays in your hands: open source, model-agnostic, infrastructure-independent, and designed to keep the most valuable part of your AI system yours.

@malte_pietsch dives deeper into this idea in his latest post:
https://t.co/8S1Sl4ROrk

[Bilge @bilgeycl]

What a search week in Berlin! 🇩🇪

On Tuesday, we hosted "Retrieve. Remember. Reason" at the @deepset_ai offices, organized with @OpenSearchProj and @cognee_, featuring Women of Search (@atitaarora 🙏), where I got to demo what's possible with a knowledge base, memory, and @Haystack_AI's orchestration capabilities. More than half the room was women, which is rare and wonderful at tech meetups! 🫶

Then yesterday, I joined the Vector Space Meetup panel at @aicampusberlin, alongside @qdrant_engine, @cognee_, @llama_index, and @n8n_io, to talk about retrieval strategies, evaluation, orchestration, and what actually goes into production. Great panel and amazing networking session! Kudos to @krotenWanderung and the team

It was a great week where I learned so much. Already looking forward to the next thing!

[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

[Bilge @bilgeycl]

We over-trust LLMs, but not worry enough about the agentic loop around them when building agents

At my @pyconit talk yesterday, I shared why agents fail and how a solid agent harness can fix them, code & slides below 👇