SlimSnap

Convert screenshots to JSON for CLI agents to read and understand

#9 in Developer tools
Developer tools
Automation
Utilities
Free
Launched 30 Oct 2026

Verified clicks, 30 days: 0
Votes: 0
Score: 14.8 pts
Comments: 0

About

SlimSnap is a screenshot tool for developers who work with CLI agents like Claude Code, Aider, and Codex CLI. Capture any area of your screen with a keyboard shortcut, annotate it with arrows and callouts, then export as structured JSON that describes every element, its text, and position. The JSON includes OCR of all labels and buttons, so your agent sees exactly what you see.

Capture runs locally with no upload or account needed. For long pages, use the scrolling capture mode to create a single JSON covering the entire page split into readable frames. You can paste the JSON directly into any agent or install a connector so Claude reads captures straight from your machine without uploading anything.

The JSON schema is published on GitHub under MIT, so you can validate it or write your own exporter. Every element gets normalized coordinates from 0 to 1, deterministic layout that stops agents from guessing where things are.

SlimSnap exports structured JSON with element positions and OCR instead of raw images, making screenshots readable in terminals and CI logs where images cannot go. Annotations carry semantic intent pointing to specific elements rather than just red pixels, so agents fix what you marked instead of guessing.

Problem: CLI agents can read your code and run tests, but cannot see images. Describing UI problems in text takes a paragraph where a screenshot would show it instantly.
Solution: SlimSnap captures screenshots as JSON with OCR and element positions, pasting into terminals and agents as text. Annotations point to specific elements so agents understand your intent.

Updated 2d ago