Quality first
In everyday work Kayts should be as good as ChatGPT and Claude, and better at research. We are not building a cheaper, worse version.
About
The people behind Kayts
Kayts is made by a small team, Monte Petrosyan and Emil Stefański, with Armenian roots and a base in Poland. We work in the open on what the product does and does not do yet.
Why we build it
Most of us already talk to an AI every day: to write, to understand something, to plan. We wanted one that works the way we do: with a photo, a PDF or a spreadsheet, in your own language, and with research you can check. Kayts answers from your sources and cites them.
We started in October 2026 as a small team. Kayts runs on a line of five models under Armenian names, from Kayts Mihr, fast and free, to Kayts Monte, our most powerful. Our own fine-tuned models come later; today we choose the best available ones and say so in the Privacy Policy.
What we hold to
Your data stays yours
Accounts and chats live in the European Union. We never sell data, and one switch decides whether your chats may help train future models.
Honest about limits
An AI can be wrong. Kayts says when its sources don't cover your question, and we publish a status page so you know when something is down.
The name
A spark, from Armenian
Kayts (կայծ) means spark in Armenian. The models carry Armenian names too: Mihr, Anahit, Siran, Nareia and Monte.