Annotake: open-source Mac tool sends only changed frames to multimodal models
Express-Phase1532 · reddit · 2026-10-30
A developer released Annotake (GPL-3.0, Mac), arguing video is a bad LLM input — so the tool captures only frames that changed, with no model in the loop. Double-tap a key, click through a task, double-tap again: you get PNGs plus a steps.md. Frames are picked by change detection using grids of grayscale tile signatures (spinners, video and blinking carets masked), and each frame gets an auto caption from the macOS Accessibility tree (e.g. Clicked button "Save"), preserving action semantics even when only images are passed. Links: annotake.app, github.com/fayzan123/annotake.
More from coding & agent
- Official Claude Code VS Code Extension Trojanized: v2.1.313 Plants SSH Backdoor — jd_pressman · 2026-10-30
- Prove2Me formalizes 2,000+ operations research papers in Lean 4 with Claude agents, 64% done — ChrSzegedy · 2026-10-30
- Optimizing product image alt text with an agentic cosine-similarity loop against Google embeddings — dejanseo · 2026-10-30
- When an AI agent retries a payment after timeout: uncertainty handling is a design decision — tallmetommy · 2026-10-30
- Gemini CLI, Jules, and Antigravity: Google's three overlapping coding tools leave users confused — finbarrtimbers · 2026-10-30
- 1-16 Coding Agents Tested: Teams Burn 36% More Tokens to Save 5 Hours — kyang8093 · 2026-10-30