v26.09-1 Release Note

Dokumentli® is a specialized Small Vision Language Model designed for European business documents. It reads, classifies, and structures documents more accurately and faster than comparable large language models.

And it runs where your documents are. Fully on-premises on a single GPU with 24 GB of VRAM, or in a private cloud. It can even operate without a network connection.

The difference compared with a general-purpose model is intentional. Dokumentli® is not designed for open-ended reasoning, coding, or general knowledge. It classifies documents and extracts specific data from them. That is precisely why it is more accurate and faster for this task, with more predictable costs.

We built it for three key groups:

• Regulated organizations (healthcare, insurance, and banking)
• European data centers looking to offer sovereign AI services
• Providers of agentic workflow software

Need Help?

Have questions or need advice or assistance? Our support team is here to help.