Show HN: Remarc – provide more contextual and structured feedback to AI agents
TL;DR
Remarc grew out of a side project problem: there was no good way to give AI coding agents contextual feedback on their own output. Plenty of tools exist for collaborating with humans, far fewer for collaborating with AI, and chat breaks down when you need to point at a specific line, a button in a screenshot, or one sentence in a paragraph. The developer describes the last ten percent of feedback and polish as the real bottleneck.
Nauti's Take
The approach is promising because feedback to AI agents really is the bottleneck: selecting and annotating beats dictating paragraphs when polishing the final ten percent. The catch is another layer sitting on top of your Mac, and the open question whether the value holds once editors and IDEs ship the same thing natively.
Worth a test for solo builders and small teams, less urgent for larger setups.
Sources