A tiny menu bar app that watches for new screenshots and instantly pastes them into your open Claude, ChatGPT, Gemini, Perplexity, or Grok tab — no clicks, no drag-and-drop.
Built specifically for macOS with native APIs — no Electron, no background daemons, no telemetry.
Screenshot appears in the AI input the moment macOS saves it. No drag, no paste.
Scans all windows on startup across Chrome, Arc, Brave, and Edge. Works with Claude, ChatGPT, Gemini, Perplexity, and Grok.
One click to lock screenshots to a chosen tab — useful when switching between apps.
Missed an auto-paste? Re-paste your most recent screenshot anytime from the menu.
One toggle installs a Launch Agent so the 📸 icon is always in your menu bar.
Pure Python + PyObjC. Sits silently in the menu bar, using almost no CPU or memory.
Two options: run directly from source (recommended for developers) or use the pre-built .app.
Clone the repo
git clone https://github.com/jonowrenn/screenshot-to-ai.git cd screenshot-to-ai
Create a virtual environment and install dependencies
python3 -m venv venv source venv/bin/activate pip install -r requirements.txt
Run the app
python app.py
The 📸 icon will appear in your menu bar.
Grant Accessibility access (required for Cmd+V fallback)
Click 📸 → Grant Accessibility and approve in System Settings.
Optional — start at login
Click 📸 → Start at Login to have it launch automatically every session.
Download & install with one command
Paste this into Terminal — it downloads, unzips, moves to Applications, and launches.
Grant Accessibility access
Click 📸 in the menu bar → Grant Accessibility → approve in System Settings.
Optional — start at login
Click 📸 → Start at Login to have it launch automatically every session.