Reliable instead of chat roulette.
Your business team assembles recurring AI workflows itself – with building blocks, without a line of code. Same input, reliably the same result.
Learn more
Build your own assistants with a visual builder, get answers from your own documents, embed chat endpoints on website and intranet, connect your systems. And because the EU AI Act applies from August 2026: risk classification and compliance documentation are built in, not retrofitted.
Same input, same result
Since 2023 we keep hearing the same two sentences from our customers: “Are we even allowed to do this under data-protection law?” and “How do we get the AI to work reliably?” An open chat window answers neither. The results differ every time, are hard to trace, and the data usually ends up in a US cloud. Fine for occasional experimenting. Not for operational use.
Your business team assembles recurring AI workflows itself – with building blocks, without a line of code. Same input, reliably the same result.
Learn moreOn-premise, hybrid or in the EU cloud. Local AI models, no silent data outflow, no training on your data. Sovereignty you can verify.
Learn moreEvery use case is automatically classified under the EU AI Act; data protection and audit evidence run along. Documentation for auditors at the push of a button.
Learn more
One platform, three operating models: fully in your own house, hybrid, or as SaaS in an EU data centre. AI models can run entirely locally.
Fully in your own data centre – for the highest sovereignty needs and regulated environments.
Sensitive data local, compute in the EU cloud – the balance of control and scale.
Quickly ready in an EU data centre – no own infrastructure, no US cloud.
Build once, use again and again: your team creates a workflow, and everyone in the company uses it. Fill in a form, read in a document, search a knowledge source, have the result checked, approve – built yourself in a visual editor.
A co-author suggests steps, fields and entire flows, and can map complete processes in just a few minutes.
Ready-made templates (e.g. meeting minutes, translate a document) and an internal catalogue holding only approved apps.
Read client files, check against internal templates – confidentiality preserved.
Pre-sort applications in a structured way, with evidence and processed sovereignly in the EU.
A consistent quote from key figures – same format, every time.
What your team builds in Habicht doesn't stay in a test window. Every assistant and every workflow can be used where the work happens: on your website, in the intranet, in existing systems.
Business users create their own assistants and mini-apps in the visual editor: forms, review steps, approvals, notifications. An AI co-author suggests the structure. Approved apps land in an internal catalogue the whole company uses.
Every assistant can be embedded as a chat: as a widget on your website, in the intranet or a customer portal. With your content as the knowledge base, source references in every answer and the access rights you define.
Via custom tools, Habicht accesses your systems in a targeted way: query databases, create reports, create records in ERP or CRM, update tickets. You decide per assistant which tools it may use.
Habicht speaks the standards your IT expects: an OpenAI-compatible API for your own applications, MCP for connecting external tools, SharePoint/OneDrive and web sources for the knowledge base. No vendor lock-in.
Habicht automatically classifies every AI use case by EU AI Act risk class and produces an exportable compliance documentation – aligned to the AI Act, GDPR, NIS2 and CRA. You prove conformity without piecing it together after the fact.
Habicht works with your own documents and shows, for every answer, what it relates to – down to the right page in the PDF. Upload files, connect websites and SharePoint/OneDrive, and control fine-grained which employee has access to which data.
The same knowledge base also feeds your embedded chats – one source, the same sourced answers everywhere.
Every answer with its source, down to the page in the PDF.
See the knowledge baseFor SMEs that want to use AI without building their own IT department. For companies that want a controlled rollout instead of shadow AI in every department. For regulated professions such as law firms, tax advisory and medicine, where confidentiality is an obligation. And for public administration, which has to work without the US cloud and needs proof.
From a quick as-a-service start through hybrid to an on-premise appliance and white-labeling for partners. We align the right scope in a call.
Quickly up and running and scaling with your company – no hardware investment required.
Sensitive data local, compute in the EU cloud – control and scale combined.
Fully in your own data centre – for the highest security needs.
Habicht under your own brand and look & feel – everything from a single source.
Transparent SaaS plans from €130/month (net); custom models on request.
“We have been developing software for Austrian ministries, NGOs and companies for more than ten years. Since 2023, everyone has been asking us the same two questions: are we even allowed to use AI, and how do we make it reliable? Habicht is our answer. No US cloud, no black box, but a platform we operate ourselves and whose answers can be verified.”
SEADEV Studios
The EU AI Act applies in full from August 2026. Organisations need demonstrable AI compliance, not declarations of intent. At the same time, per Bitkom 78% of companies want less dependence on US cloud providers, and public bodies are urged to prefer open, European solutions. Habicht is built for exactly this environment.
The EU AI Act applies in full.
Source: Bitkom survey on cloud sovereignty · EU AI Act, obligations entering into force from 08/2026.

Sovereignty, compliance and operations – answered briefly and concretely.