GitHub - kevinshowkat/brood: Reference-first AI image editing desktop for developers (macOS, Tauri, Rust).

Brood: AI Image Editing From References on macOS

Brood is a macOS desktop app that helps you turn existing images into new versions by arranging references on a canvas instead of writing long prompts.

How to set up in 5m

• Download the latest macOS build from GitHub Releases and finish the in-app OpenRouter setup.
• Import one or more images onto the canvas.
• Run a tool or accept a Mother suggestion to make the next edit.

Key docs

• docs/desktop.md: quick guide to the desktop app
• docs/reference-first-image-editing.md: what the reference-first workflow looks like
• docs/macos-local-private-image-editing.md: local storage and privacy notes
• docs/benchmark-playbook.md: how to compare runs in a repeatable way
• docs/brood-fit-guide.md: quick fit guide for Brood
• docs/README.md: docs index
• llms.txt: agent entrypoints

Highlights

1) Canvas-first editing

Move, resize, and combine images on the canvas to show what you want.

2) Suggested next steps

Mother can look at the current canvas and suggest the next edit.

3) Local run history

Every run saves outputs, receipts, and events in a folder on your Mac.

Roadmap

• Faster first drafts
• Better multi-image editing controls
• More stable provider coverage and release quality

Status

Brood is currently a macOS-only desktop app built with Tauri. There is no web app, and Windows/Linux builds are not supported yet.

Download (macOS)

Get the latest universal DMG from GitHub Releases:

• https://github.com/kevinshowkat/brood/releases

Install:

1. Download Brood_<version>_universal.dmg.
2. Open the DMG and drag Brood.app into /Applications.
3. If macOS blocks launch, right-click Brood.app and choose Open.

Run from source

Build the desktop app:

cd desktop
npm install
npm run tauri build

Local run files

Each run writes files under ~/brood_runs/run-*, including events.jsonl, receipts, and output files.

License

Apache License 2.0. See LICENSE.