— but even when it's wrong, it gives the developer a starting point instead of a cold start. I'm not sure if this is genuinely useful or a gimmick, and I'd love honest takes.

Free tier is 20 reports/month, no credit card. The link goes straight to a demo you can break things on without signing up.

Things I'd love feedback on:

- Does Execute Fix feel useful or like AI-feature-creep?

- For those who've used BugHerd / Marker.io / Usersnap — what's the friction that made you bounce or stay?

I'll be in the thread for the next ~12 hours. Ask me anything about how it's built or why I made specific calls.

1f:T736,

I've been building software for clients and SaaS products for a while, and the same pattern kept showing up: a user or stakeholder reports a bug, the developer asks three follow-up questions ("which browser? what screen? can you screenshot?"), then spends 20 minutes reproducing it before they can even start fixing. Half the time the original report was so vague the dev gave up and waited for it to happen again.

So I built Feedzap. It's a script tag (~one line) you drop into any web product. Users click a button, point at the broken thing, type one sentence. The report that lands in your Linear/Jira/Slack includes the screenshot, the CSS selector of the element they clicked, the full URL, viewport, browser, and OS — automatically. No follow-ups needed.

The part I'm most interested in HN's opinion on: Execute Fix. Instead of just routing the report as a ticket, it takes the captured context (selector, screenshot, description) and generates a code patch you can review before applying. It's right ~70% of the time on CSS/copy bugs from my testing, much less on logic bugs — but even when it's wrong, it gives the developer a starting point instead of a cold start. I'm not sure if this is genuinely useful or a gimmick, and I'd love honest takes.

Feedzap

About