✨Return to a more utilitarian design:
We got rid of unnecessary visual elements and returned to a style similar to v1.
Now the interface has fewer distracting details, a higher density of useful information, and a focus on neural network responses.
🎹 Hotkeys for selecting models:
In the settings, you can now assign up to 5 hotkeys for quickly switching between neural network models.
This allows you to instantly change the model according to the type of question (algorithms, system design, behavioral).
💬 Chat:
It is now possible to allow models to search for information on the internet. This provides access to relevant context that is not included in the training sample, as well as more accurate answers on the latest technologies and libraries.
Tools are now called in parallel. The result: graphs are built faster in system design mode, and the interface responds noticeably faster.
🔧 Fixes:
- The Tools block can be collapsed, similar to audio transcriptions.
- Fixed a bug where a group could collapse after deleting an element.
- On macOS, when the window is hidden and a screenshot of the selection area is taken, the selection area is now displayed correctly.
