Nitpicks

Point Nitpicks at a page, record your screen and voice, and describe what you want changed, and the agent will read the context of the page itself to figure out which elements your recording actually refers to. The AI frontend bug fixer interprets the recording and writes the code that addresses it. It's the fastest path when a problem is easier to show than to type out. No ticket required.

Press Ctrl + Shift + A to enter annotation mode, then click each element you want to flag and write what should change. This suits a structured QA review, where you walk a page and log issues one by one. The annotation to code loop stays tight, because each note maps to a specific on-screen target. The agent knows exactly what to touch.

Once feedback is captured, Nitpicks implements the change automatically and sends it to your dev team for review within minutes, so the fix is often ready before the person who spotted the bug has moved on to their next task. Nothing skips code review. You still approve the result through your normal workflow. That's it.

The agent plugs into the tooling you already use. It runs your test suites, applies your linters and formatters, executes custom CLI commands, and verifies that changes compile in your build pipeline. Every configured check runs before a change reaches review, so broken code gets caught early. Wrap it around your workflow, not the other way around. That means your existing quality bar still applies to anything the agent writes, and you don't have to loosen your standards just to move faster.

Anyone who can record a video can contribute to product development, not just developers. Product, QA, design, and even sales can flag improvements and get code delivered, which spreads the load beyond a single engineering team. That's the point. If you can talk through a problem, you can help fix it, and the tool handles the rest of the translation into code.