







Château abandonné
Je suis allé voir ce château l'année dernière, en y allant j'ai croisé des chars et des militaires en exercice, assez étonnant comme ambiance...








Je suis allé voir ce château l'année dernière, en y allant j'ai croisé des chars et des militaires en exercice, assez étonnant comme ambiance...
Full disclosure: I'm the developer, and Pinpoint is free and open source (MIT) — no paid tier, no tracking, nothing leaves your Mac.
The problem I built it for: I pair-program with AI agents (Claude Code, Codex) and kept wasting time pasting a screenshot, then typing a paragraph to explain which button or icon I meant. Worse, a lot of chat UIs keep only the image and drop the text you copied with it.
What Pinpoint does: hit a global shortcut, drag a region, drop numbered markers on what matters, add a short note to each, and copy. Your clipboard gets the annotated image plus a structured text that maps every marker to a position and description — so the agent knows exactly what you point at.
Mac-native details this community tends to care about:
Download + source: https://github.com/croustibat/Pinpoint
I built it for myself and use it daily. Feedback very welcome — especially on the generated prompt format. What does your screenshot-to-agent workflow look like today?