v1.1.8 is out, and it's centered around two things: a new way to work with files in AI conversations, and a big pass on rendering performance for canvas and UI elements.
You can now bring files directly into your AI conversations — the AI reads and works with the file content as part of the chat, so you can ask questions or get help based on what's inside, without extra steps.
For example, once you have an image and a multimodal model configured, you can use it to recreate that image — a simple way to see the new file support in action.
A large chunk of this release went into making element rendering smoother, especially in the canvas and whiteboard:
Smoother canvas zoom/pan on macOS
