Features

• Floating file explorer when the sidebar is narrow for easier navigation.
• Tall tool calls (e.g. file edits, shell commands) are collapsed by default and can be expanded.
• File explorer only opens automatically on wider screens to avoid clutter on small layouts.
• Line count shown in names for large pasted text files.
• Consolidated toolbar and sidebar into one web contents view.

Bug Fixes

• Narrow settings screen layout and overflow handled better.
• Minimum window size set so the UI stays usable.
• Empty reasoning blocks are hidden.
• Window resizing is now more visually responsive.
• Close-tab animation is consistent.
• Rapid sidebar toggle clicks work reliably.
• App no longer flickers when swapping files.
• AI Gateway supports model IDs with version-style suffixes (e.g. grok-4-1-fast).

🧼

36 commits since v1.13.0

More will follow.

Features

• File panel and explorer - A new file explorer lets you view all the files in a project. File panel allows viewing files larger in place of viewing the app.
• Hideable app panel - The app preview panel can now be toggled, giving you more space when you need it.
• Context menu and file actions - Right-click context menus and a unified file actions menu make it easier to open, copy, and manage files.
• Upgraded model heuristics - Detection and handling of newer Claude models improved.

Bug Fixes

• Fixed overlap between project name and controls in the header.
• Images are no longer incorrectly rounded in previews.
• All models are now correctly available on paid plans.
• Windows drive letters on attached folders now display correctly in chat.
• Absolute paths now work correctly in glob and copy-to-project operations.
• Windows-made projects can now be imported on macOS.
• Fixed agent clarity around read-file returning a directory listing.
• Fixed truncation of ls output.
• Avoid repeating already-copied files.
• Uploaded files section no longer appears when empty.

🍀

69 (👌) commits since v1.12.0

More will follow.

Features

• Source images for image generation - The agent can now use source images during generation. Uploaded or previously generated!

• Together AI and DeepInfra image support - Both providers now work for image generation.
• Virtualized project sidebar - Large project lists scroll smoothly; favorites sorted by recent activity.
• Improved project toolbar layout, better use of space.

Bug Fixes

• Support for additional image formats (e.g. JPEG) when copying.
• Folder picker switched to faster, Electron-native implementation.
• AI provider suggestion only shown on correct error type.
• Agent better guided away from direct access to attached folders.

🌄

29 commits since v1.11.0
More will follow.

Bug Fixes

• Improved suggestions for adding an AI provider when the right error type is detected.
• Agent now encouraged to try more approaches or ask the user when encountering unsupported file formats.
• Folder picker switched to a faster, Electron-native implementation.

📂

4 commits since v1.11.0
More will follow.

Features

• Folder attachments - Attach local folders to your project.

• Output files are now grouped by file type in the files panel.
• Favorite projects are surfaced in the command menu.
• Minimax models are now recommended.

Bug Fixes

• Copy, "View as Code", and copy HTML now work correctly for Markdown and HTML files in the file viewer.
• Handle display unicode in filenames properly (e.g. macOS screenshots).

📁

93 commits since v1.10.0
More will follow.

Features

• Web search finally. The agent can search the web! Works with many AI providers and your free Quests account, of course.
• Prev/next tab and undo tab close shortcuts.

Bug Fixes

• I fixed the bugs I introduced as I built web search.

🔎

Sorry about the whole ".10" thing. It's an ugly practice programmers do called, "pedantic versioning." 2.0 will be something more special.

28 commits since v1.9.2
More will follow.

Bug Fixes

• AI Gateway: Allow Claude Opus 4.6 in heuristic tagging.
• Studio: Stop showing tool calls that never executed.

🤏

3 commits since v1.9.1

More will follow.

Features

• Larger file view – Single artifacts (e.g. images, files) now use a bigger, clearer display.

Bug Fixes

• Paths for generate image and read file suggestions are normalized.
• HTML code can be viewed again in the app.
• Consistent background colors on load to prevent flicker.
• CSP updated so font-src is allowed and fonts load correctly.
• Old low-credit error messages are hidden when no longer relevant.
• AI gateway only logs model fetch errors once to cut down noise.

🧰

20 commits since v1.9.0

More will follow.

Features

• Image generation: Ask the agent and it can make images for you using many different providers. Quests, OpenRouter, OpenAI, Google, and more are supported.

Bug Fixes

• Upgraded to AI SDK v6 and fixed some provider incompatibilities.
• External links now open safely while regular navigation still works as expected.

🖼️

42 commits since v1.8.2

More will follow.

Bug Fixes

• Manual download link now opens in browser.
• Eliminated some console warnings on update.
• Deep linking into the app now works on Linux.

🐡

13 commits since v1.8.1

More to come.