Captures a screenshot, lets you mark it up with arrows and callouts, and copies it as JSON — element bounding boxes, text pulled by local OCR, and each annotation exported as structured intent — so terminal-only agents like Claude Code or Aider can read a UI they can't see as an image. Handles whole scrolling pages too, split into frames under one JSON capture. Capture and OCR run entirely on your machine, and the JSON schema is published under MIT, though the capture apps themselves stay closed-source.
What holds up
- +OCR and capture run fully locally — screenshots never leave your machine
- +JSON schema is open under MIT — read it, validate it, or write your own exporter
- +Annotations export as structured intent, not just colored arrow pixels
Mind the limits
- −Capture apps are closed-source — only the schema and Claude Code skill are open
- −Only useful for text-based agent workflows — just a screenshot tool otherwise
Picked here? Add the badge
Put it on your site. It links back to this page. Free for every listed product.
<a href="https://stillworks.dev/products/p/slimsnap/"><img src="https://stillworks.dev/badge/slimsnap.svg" alt="Picked by StillWorks" width="230" height="54"></a>