Runs entirely on your Mac.
AI-powered color accessibility analysis for macOS.
ColorSense lives in your menu bar. Press a shortcut, draw a region anywhere on your screen, and instantly see how it looks to someone with a color vision deficiency — and a corrected version that everyone can distinguish.
- One shortcut, any app — Press ⌘⇧Y from anywhere to start a capture
- Side-by-side comparison — Original image next to a CVD-corrected version
- AI analysis — Powered by Google Gemini; identifies ambiguous color pairs and explains the issue in plain English
- CVD-safe color correction — Remaps problematic colors using the Wong (2011) scientifically validated palette, distinguishable under deuteranopia, protanopia, and tritanopia
- Save the fixed image — Export the corrected image as PNG or JPEG
- Works everywhere — Normal windows, Safari, Chrome, full-screen apps, and apps in separate Spaces
- Configurable model — Choose any Gemini model from the menu bar
- macOS 14 (Sonoma) or later
- A free Google Gemini API key
- Screen Recording permission (macOS will prompt on first use)
- Go to the Latest Release
- Download
ColorSense-1.0.0.zip - Unzip and drag ColorSense.app to your
/Applicationsfolder
First launch — Gatekeeper notice:
Warning
ColorSense is not notarized. On first launch macOS will block it — right-click (or Control-click) ColorSense.app → Open → click Open in the dialog. macOS only asks this once.
Requires Xcode 15+ and macOS 14+.
git clone https://github.com/WW-Web-Infra/ColorSense.git
cd ColorSense
open ColorSense/ColorSense.xcodeprojIn Xcode:
- Select your signing team under Signing & Capabilities
- Press ⌘R to build and run
On first launch, ColorSense walks you through a short setup:
- API Key — Paste your free Gemini API key from Google AI Studio. Select a model (default:
gemini-2.5-flash) - Screen Recording — Grant permission so ColorSense can capture your screen region
- Done — You're ready to go
You can update your API key or model anytime via the menu bar icon → Set API Key…
| Action | How |
|---|---|
| Capture a region | Press ⌘⇧Y or click the eyedropper icon in the menu bar |
| Draw the region | Click and drag over any part of your screen |
| Cancel | Press Esc |
| Save the fixed image | Click Save ColorSense Image… in the result window |
| Change API key / model | Menu bar icon → Set API Key… |
| Re-run setup | Menu bar icon → Reset Setup… |
- You draw a selection rectangle over any screen content
- ColorSense captures that region using ScreenCaptureKit
- The image is sent to Gemini AI, which identifies color pairs that are ambiguous for people with color vision deficiencies
- The app remaps those colors to the Wong (2011) CVD-safe palette — seven colors proven distinguishable under all common CVD types
- The result panel shows the original and corrected images side by side, with a plain-English explanation and a list of the affected color pairs
- Images are sent to Google's Gemini API for analysis. Google's API usage policies apply.
- Your API key is stored securely in the macOS Keychain — never in plain text or in any file
- No data is collected or stored by ColorSense itself
| Permission | Why |
|---|---|
| Screen Recording | To capture the region you select |
| Network (outbound) | To send the image to the Gemini API and receive the analysis |
| Accessibility | To register the global ⌘⇧Y keyboard shortcut |
- Color correction palette: Wong, B. (2011). Points of view: Color blindness. Nature Methods, 8, 441
- Screen capture: Apple ScreenCaptureKit
- AI analysis: Google Gemini
PRs are welcome! Bug fixes, new features, and palette improvements are all fair game. Open an issue first for larger changes so we can align on direction before you invest the time.
MIT License — see LICENSE for details.