AI for the public sector — data that stays in Sweden
OpenAI-compatible API on Swedish hardware. GDPR-compliant by design. All inference in Sweden.
Why municipalities can't use ChatGPT
Data transfers to the US
GDPR Article 46 and the Schrems II ruling mean personal data cannot be transferred to the US without adequate safeguards. OpenAI processes data on American servers.
Inadequate data processing agreements
Standard terms from big tech companies are not sufficient for sensitive operational data in the public sector. Public organizations handle data about citizens, students, and patients.
Vendor lock-in & costs
Price changes, proprietary formats, and dependence on a single vendor create risks for public organizations with limited budgets.
Staik solves the public sector's AI problem
- All data is processed on servers in Sweden
- All data processed on our own hardware — never via foreign cloud services
- OpenAI-compatible — works with existing tools and integrations
- Transparent pricing per organization, not per token
How it works technically
Tailored for the public sector
Municipalities & regions
Administrative support, document management, citizen dialogue
50+ municipalities are exploring AI solutions — be early
Schools & education
Teacher support, lesson planning, administrative relief
GDPR-safe for student and staff data
Healthcare
Documentation support, journal summaries, administrative processes
Meets requirements under GDPR and the Patient Data Act
Simple, predictable pricing for the public sector
All prices excl. VAT. Billed monthly or annually.
Pilot
Free for 30 days
Up to 5 users, email support
Enterprise
Custom pricing
Dedicated infrastructure, custom solution, phone support
Frequently asked questions
All data is processed and stored on servers physically located in Sweden. We never use cloud services outside the EU/EEA for data processing.
For contracts under the direct procurement threshold (700,000 SEK excl. VAT), you can sign agreements directly. We help you with the necessary documentation.
No. Your data is never used for training or improving models. All inference happens in real-time without storing prompt content.
We never store prompt content or results after the response is delivered. All inference runs on our own hardware in Sweden without third-party access.
Yes. Staik API is OpenAI-compatible, which means systems already integrating with ChatGPT or GPT-4 can switch to Staik without code changes.
Ready to try? Start with a free pilot.
Contact usWe respond within one business day. No commitment required.