Free
0 € forever
- About 100 conversations a month
- 200 pages indexed
- "Powered by Kaval.AI" in the chat
Don't hesitate to ask for a DEMO! Basic plan is FREE!
Our robots index every page of your website, discovers products, services, prices, FAQs and talks to your website visitors.
0 € forever
29 € a month, excl. VAT
89 € a month, excl. VAT
249 € a month, excl. VAT
Let's talk free consultation
Prices exclude VAT and yearly billing gives two months free. Create an account to compare the plans in detail, or about Enterprise.
The website chatbot is what we sell off the shelf; for larger organisations we build AI by hand, on your own data, your own servers and your own workflows.
The website chatbot and most of our client work run on kavalai, our open-source Python library for agentic workflows. Agents are declared as typed workflow graphs rather than hand-written glue code, and the chat widget on this very page is our own hosted chatbot, built on it. The same building blocks go into client projects, so nothing we deliver depends on a framework we cannot fix ourselves.
Inputs, outputs and tool calls are typed models, so a workflow has a contract to test against. An evaluation suite replays real conversations before every deploy, and a failing case leads straight back to the run that produced it.
The model's output is validated before your code sees it. Streaming, retries, rate limiting and sessions are handled by the library, and any provider works behind one interface, so switching models is a configuration change.
Every session, run, node and model call is recorded and browsable. Conversations lead to runs, runs to tasks, and each task opens into its input and output, so a bad answer can be traced to the node that caused it.

Timo Petmanson is a data scientist and engineer with over 15 years of experience in data analytics, natural language processing and data engineering.
He studied computer science at the University of Tartu, was a junior researcher working with Estonian e-health data and a major contributor to the Estonian natural language processing toolkit ESTNLTK.
He later worked as software engineer at Google in Zürich in ads team, was a data scientist at Venturebeam in Tallinn and developed datalakes at Veriff.
In 2026 he founded Kaval.AI to build AI workflows, chatbots and tools for companies.
Ask about the Enterprise plan, tell us what you are building, or what is taking up your team's time, and we will get back to you.