Introducing Kai: Your AI Data Engineer, Now in Public Beta
Kai is an embedded AI assistant - a context-aware data engineering agent that understands your data, transformations, flows, and configurations, and can help you explore, build and debug (with your approval)
We're excited to announce that Kai, Keboola's embedded AI assistant, is now available in Public Beta for all users. Kai isn't just another chatbot—it's a context-aware AI agent that lives inside your Keboola project and works alongside your team as a full-fledged data engineer. Unlike generic AI tools that require you to copy-paste error logs and guess at solutions, Kai already knows your flows, lineage, schemas, and logs. Just describe what you need in plain English, and Kai handles the rest—from building pipelines to debugging failed jobs.
What Kai Can Do
- Build End-to-End Pipelines: Describe your data flow in natural language, and Kai configures extractors, writes transformations, and sets up writers—with your approval at each step.
- Debug Failing Jobs: When something breaks, Kai analyzes error logs, traces the root cause through your lineage, and proposes targeted fixes you can test safely in branches.
- Write SQL & Python: Generate production-ready transformations without writing code.
- Auto-Generate Documentation: Kai scans your flows and components, and storage objects to create comprehensive documentation stored in descriptions.
- Answer Questions About Your Data: Ask in plain English. Kai writes the SQL, runs it, and shows the results.
- Create Data Apps: Describe the dashboard you want, and Kai generates the Streamlit code.
Getting Started
- Open any Keboola project
- Click the Kai icon in the header
- Start with: "What does this project do?" or "Why did my last job fail?"
We're actively gathering feedback during Public Beta—use the thumbs up/down buttons in the chat to share your experience.
To learn more, see our docs.