Show HN: We built an AI to review your pull requests

by gentios • 1 year ago

53 5 1 year ago

Hey HN,

We’re two developers (co-founders) with a team of 20 who got tired of spending hours reviewing PRs, so we built Infinitcode.ai, an AI-powered code reviewer that:

- *Summarizes PRs in plain English*: No more deciphering 1,000-line diff jungles

- *Catches more than bugs*: Security holes, performance pitfalls, code smells, even typos (yes, we’ll flag “vurnerabilities” and vulnerabilities)

- *Zero onboarding*: Works instantly—no “let me learn your codebase for weeks” nonsense.

Why we’re posting: We’re in alpha and need brutal honesty. Roast our tool, mock our UI, or tell us why AI will never replace your team’s Senior Engineer.

Free alpha access for 100 teams: All we ask is feedback.

Give it a try: https://infinitcode.ai

Demo repo: https://github.com/infinitcodecom/infinitcode-ai-demo

zeeroman• 1 year ago

This is interesting. I’ve been skeptical about AI code review tools, mostly because the ones I’ve tried either spam low-value suggestions or miss obvious context. But I gave Infinitcode a try on a recent PR (Python/Django project), and it actually gave a pretty solid summary and caught a couple things I missed on first pass — including an inefficient query I hadn’t optimized yet.

Still early, and I wouldn’t fully trust it on critical code, but as a “reviewing assistant” it’s already saving me time. Curious how it holds up on larger or more domain-specific codebases. Definitely worth keeping an eye on.

gentios2• 1 year ago

Thank you for your honest feedback and giving it a try. Before launching to public we used it internally since we are a dev team (still using it :D) and we have tried it into specific domains such as: fintech, e-commerce etc. with larger code bases and have really performed well

The purpose of it has always been to be a good assistant of the code reviewer it self by giving a clean summary and suggesting fixes

Of course we work in new features that will be quite helpful for developers

vegims• 1 year ago

Just wanted to say, really loving this so far. Tried it on a couple of real PRs today and the summaries alone saved me a chunk of time. It actually understood what the change was doing instead of just paraphrasing file diffs, which is more than I can say for most tools I’ve tried.

Also appreciated that it flagged a subtle performance issue I almost missed. This feels like the kind of assistant I wish GitHub had built into reviews by default.

tuntuntunsahur• 1 year ago

Cool idea! Summarizing PRs and spotting issues sounds useful, but how does it deal with tricky code that needs some context? Curious how it works on real projects.

gentios2• 1 year ago

We fetch file changes from github (we fetch only, we don't store code) and based on that we build the context.

In order for the AI to be more precise you can as well fine tune from the dashboard it self such as:

- multiple models - you can set the tone - some prompt engineering

you can as well add code styling for: functions, variables etc.

Suggestions are welcomed :)

Show HN: We built an AI to review your pull requests

Related Stories