Changelog

• You can now add multiple source images per canvas. Great for grouping related work together.
• You can now use image reference inputs when editing an image.
• Added support for Grok Imagine 2.0 model.
• Improved command palette performance for bigger workspaces.
• Fixed the focus outline in the zoom preview.

• You can now perform precise image edits on specific tagged regions when working with branch edits. This experimental feature allows you to annotate different parts of the image and then send it to the model in one go to edit everything.
• The command palette now opens faster.
• Settings have been improved with new tabs and a faster provider selection process

• You can now use the Qwen Image provider
• The menu bar now opens the window with a single click, and right-clicking opens a context menu.
• Fixed issues with FAL upscaling cost estimates.

• You can now upscale images using the local RealESRGAN model or upscale models from the cloud provider FAL.
• You can now compare image versions using a slider between two versions.
• Added shortcut to copy the image image of a version node. Use the ⌘ + C keyboard shortcut, right click on the node and select copy image, or with a node select, use the command palette to copy the image

• Support macOS 15+, Liquid Glass now gated to macOS 26+

• Fix canvas lag with many images
• Add "Current" version marker to changelog

• You can now use the command palette(⌘ + K). Great for keyboard the power users.
• Added batch mode canvas to process multiple images at same time.

• Fixed an issue where the canvas recenter offset was incorrect

• The macOS app and help pages have been localized.

• Stop mouse version node clicks recentering canvas

• Add local MCP server for integration with AI agents
• Add experimental feature toggles, keep the app minimal.
• Add canvas node order label
• Support text generation from blank nodes

• You can now use Google's latest model, the Gemini Flash Lite image model.
• You can now navigate the image preview using WASD keys.
• You can now see estimated model durations based on your own usage.

• You can now use the OpenAI provider with GPT Image 2

Fixed issues with canvas node navigation

• You can now use keyboard pan and zoom controls in the full image preview.
• You can now add watermarks to your images.
• Model pickers now show the estimated cost per image.

• You can now navigate the canvas using your keyboard.
• Fixed an issue where image overlay scrolling and zooming were interfering with canvas gestures.

• You can now track and view the cost associated with each image generation edit and the total cost for the canvas.
• We have improved the actions available for graph nodes.
• You can now see the live elapsed time for edits within nodes and in the details panel.

• You can now use local image background removal.
• You can now share your thoughts with the developers from inside the app.

• You can now use native focus rings for a cleaner interface.

babblebrush v0.2.0

babblebrush v0.1.0