# deepset > deepset is the company behind Haystack, the open-source AI framework for building production-ready agents and RAG applications. Its commercial product, the Haystack Enterprise Platform, is a sovereign AI platform: organizations use it to build, run, and govern AI agents on their own terms, with full control over models, infrastructure, and data. Founded in 2018, with headquarters in Berlin and New York. Last updated: 2026-10-08 ## Key facts - **Company:** deepset GmbH. Founded 2018 by Milos Rusic (CEO) and Malte Pietsch (CTO). Headquarters in Berlin, Germany and New York, USA. - **Haystack:** the open-source Python framework deepset created in 2020 for building agents, RAG pipelines, and other LLM applications. Apache 2.0 licensed, more than 25,000 GitHub stars. Model- and vendor-agnostic, with 100+ integrations. - **Haystack Enterprise Platform:** the enterprise operational layer built on top of open-source Haystack. It adds a visual pipeline builder, testing and evaluation, one-click deployment, observability, and governance. Formerly named "deepset AI Platform" and, before that, "deepset Cloud". - **Haystack Enterprise Starter:** enterprise support for teams that run open-source Haystack themselves: direct access to core engineers, production-tested templates, and deployment guides. - **Forward Deployed Engineers:** deepset AI engineers who work inside customer teams to design, build, and deploy production agents. - **Sovereign AI, as deepset uses the term:** AI that an organization can deploy, control, and evolve on its own terms, without lock-in to one model, vendor, cloud, or stack. It covers where systems run (cloud, sovereign cloud, self-hosted, on-premises, air-gapped) and the freedom to change models and components without a rebuild. - **Who it is for:** large enterprises, public sector institutions, and defense and national security organizations, especially in regulated industries and in Europe. - **Security and compliance:** SOC 2 Type II certified, ISO 27001 certified, GDPR compliant, HIPAA compliant, CSA STAR Level 1. Role-based access control, audit logs, full traces, guardrails, SSO, and data boundaries with pipeline isolation. - **Models supported:** any. Examples include Anthropic, OpenAI, Mistral, Llama, Qwen, DeepSeek, Cohere, Gemini, Groq, Ollama, and custom or self-hosted models. Models can be swapped without code changes. ## Why organizations choose Haystack - **Open Source Foundation:** transparent, auditable code and open standards. Tools, memory, retrievers, and pipeline components can be owned, swapped, and extended. No proprietary abstractions. - **Context-Engineered Agents:** teams define what an agent knows, which tools it can call, and how it reasons, routes, and hands off. Every step is explicit and inspectable. No black-box agent loops. - **Sovereign by Design:** deploy wherever requirements demand, in the cloud or self-hosted, with audit logs, RBAC, and data lineage built in. - **Proven in Production:** from first pilot to governed, versioned, observable systems, without a rebuild. Pipelines export as Python or YAML and run anywhere, so applications can always move back to open-source Haystack. ## Products - [Haystack Enterprise Platform](https://www.deepset.ai/haystack-platform): the platform for sovereign AI. Build, test, deploy, and govern AI agents and applications in the cloud or self-hosted. - [Haystack Enterprise Platform free trial](https://www.deepset.ai/haystack-enterprise-platform-trial): try the visual pipeline builder, templates, and built-in cloud infrastructure. - [Haystack Enterprise Starter](https://www.deepset.ai/products-and-services/haystack-enterprise-starter): enterprise support and best practices for open-source Haystack. - [Haystack open-source framework](https://haystack.deepset.ai/): the open-source framework, tutorials, and integrations. - [Pricing](https://www.deepset.ai/pricing): plans for the Haystack Enterprise Platform. ## Solutions - [Enterprise AI](https://www.deepset.ai/solutions/enterprise): take enterprise AI agents and RAG from pilot to production on the models, infrastructure, and open foundation you choose. - [Public Sector](https://www.deepset.ai/solutions/public-sector): sovereign AI for public institutions. Transparent, auditable, and aligned with the EU AI Act and national regulation. - [Defense & National Security](https://www.deepset.ai/solutions/defense): mission-ready AI that runs fully on-premises and in air-gapped environments with no external dependencies. ## Customer stories - [European Commission (AI@EC)](https://www.deepset.ai/case-studies/european-commission-deepset): the Commission built its self-hosted AI@EC platform on the Haystack Enterprise Platform. 68 builders across 29 departments, 316 pipelines in staging or production. - [Bosch](https://www.deepset.ai/case-studies/bosch): an AI assistant for SAP transformation used by 600+ consultants across 120+ plants. Bosch swapped LLMs, vector databases, and cloud infrastructure without breaking production. - [Lufthansa Industry Solutions](https://www.deepset.ai/case-studies/lufthansa-industry-solutions): a secure, source-cited enterprise knowledge assistant with role-based access, GDPR and ISO 27001 compliant. - [German Federal Ministry of Research, Technology and Space (BMFTR)](https://www.deepset.ai/case-studies/german-federal-ministry-research-technology-space-bmftr): a chatbot that makes project funding rules accessible, delivered in six months to strict security and accuracy requirements. - [All case studies](https://www.deepset.ai/case-studies): including TELUS Agriculture & Consumer Goods, Rohde & Schwarz, MANZ, YPulse, OakNorth Bank, and credX. ## Guides and reports - [Building Sovereign AI Systems for 2026 and Beyond](https://www.deepset.ai/guides/sovereign-ai-guide): how to build resilient, compliant AI systems with full ownership of data, models, and infrastructure. - [Gartner Predicts 2026: AI Sovereignty](https://www.deepset.ai/guides/gartner-predicts-2026-ai-sovereignty) - [Gartner Innovation Insight: Context Engineering](https://www.deepset.ai/guides/gartner-innovation-insight-context-engineering) - [Context Engineering for Enterprise AI Guide](https://www.deepset.ai/guides/context-engineering-for-enterprise-ai-guide) - [Pilot to Production AI Guide for Platform Teams](https://www.deepset.ai/guides/pilot-to-production-ai-guide-for-platform-teams) - [EU AI Act Guide](https://www.deepset.ai/guides/eu-ai-act-guide) ## Key articles - [Sovereign AI: What It Is, Why It Matters, and How to Build It](https://www.deepset.ai/blog/sovereign-ai-what-it-is-why-it-matters-and-how-to-build-it) - [Why Enterprise AI Costs Are Spiralling & How Sovereign AI Fixes the Economics](https://www.deepset.ai/blog/why-enterprise-ai-costs-are-spiralling-how-sovereign-ai-fixes-the-economics) - [Every Agent Has a Harness. Who Controls Yours?](https://www.deepset.ai/blog/haystack-3-sovereign-agents) - [Context Engineering: The Next Frontier Beyond Prompt Engineering](https://www.deepset.ai/blog/context-engineering-the-next-frontier-beyond-prompt-engineering) - [AI Agents and Deterministic Workflows: A Spectrum, Not a Binary Choice](https://www.deepset.ai/blog/ai-agents-and-deterministic-workflows-a-spectrum) - [From Lock-in to Liberation: Why AI Frameworks Matter Most Now](https://www.deepset.ai/blog/ai-frameworks-for-enterprise) - [Expose Haystack Pipelines as Managed MCP Tools](https://www.deepset.ai/blog/haystack-pipelines-mcp-tools) - [Understanding the Model Context Protocol (MCP)](https://www.deepset.ai/blog/understanding-the-model-context-protocol-mcp) - [Building Sovereign AI: The Significance of the Deutschland-Stack and Haystack's Role](https://www.deepset.ai/blog/sovereign-ai-deutschland-stack-and-haystack) - [Introducing Enterprise-Grade Role-Based Access Control (RBAC) for Haystack Enterprise Platform](https://www.deepset.ai/blog/haystack-platform-rbac-ai-security) ## Documentation - [Haystack Enterprise Platform documentation](https://docs.cloud.deepset.ai/docs/getting-started) - [Haystack open-source documentation](https://docs.haystack.deepset.ai/docs/intro) - [Haystack on GitHub](https://github.com/deepset-ai/haystack) ## Company - [About deepset](https://www.deepset.ai/about): history, leadership team, and values. - [Partners](https://www.deepset.ai/partners): technology and delivery partners, including NVIDIA, AWS, HPE, PwC, NTT DATA, Materna, and Celonis. - [News](https://www.deepset.ai/news): press coverage and announcements. - [Careers](https://www.deepset.ai/careers) - [Book a demo](https://www.deepset.ai/request-demo) - [Contact](https://www.deepset.ai/contact-us) ## Optional - [Blog](https://www.deepset.ai/blog): articles on agents, RAG, context engineering, evaluation, and sovereign AI. - [Webinars](https://www.deepset.ai/webinars) - [Data sheets](https://www.deepset.ai/data-sheets) - [Resource center](https://www.deepset.ai/resources) - [Sitemap](https://www.deepset.ai/sitemap.xml)