July 1, 2025 - More Models, More Next Steps

Choose from the best of the best models from OpenAI, Anthropic, and Google

Mito users can now select between OpenAI, Claude, and Gemini models. All without setting an API key. Switch back and forth in the same chat to experiment with different responses.

After finishing its work, the agent will now suggest next steps for your analysis.

After each response from the agent, you can revert changes to go back to the previous state. This should give you the confidence to explore with the agent, knowing you have a previous version of your work you can easily go back to.

Fix bug that causes the notebook footer to sometimes not show up.
Clicking on a citation now always scrolls the cell into view, even if the cell is far away.
Citations can now reference multiple lines in a code cell, making the highlighted block of code much more useful.
When the user is scrolled up in the chat and a new message comes in, we no longer auto scroll back to the bottom. This made is really challenging for users to follow along with the agent as it was working.
Improves the chat so that it is more likely to always respond with the full code cell, even in very long code cells.
Try to help the agent respond better to simple inputs like "Hi" better without it ending up in an infinite loop of writing markdown cell.
Makes databases more discoverable by adding the "Connect Database" button to the blank taskpane screen.

Last updated 18 days ago

Was this helpful?