Codeball

AI CODE REVIEW – Wait less for review, save time and money

42 followers

AI CODE REVIEW – Wait less for review, save time and money

42 followers

•

•

Codeball approves Pull Requests that a human would have approved. This means less waiting for review. Less rubber stamping also means the team gets to spend more time on the trickier PRs.

Free Options

Launch tags:Developer Tools•Artificial Intelligence•GitHub

Launch Team

Render Sandboxes: a computer for every agentRun untrusted code safely. Idle sandboxes cost nothing

Promoted

Scrimba

We've been trying out Codeball for a few weeks now at Kitemaker and I have to say so far it's been remarkably accurate. We haven't seen a single false positive where it flags a PR as approved where we wouldn't have approved it ourselves manually.

Report

4yr ago

Polar

Maker

Thanks @ksimons, it's been awesome to see Codeball doing it's magic for Kitemaker!

Report

4yr ago

GitButler

Maker

Thanks @ksimons 🚀

Report

4yr ago

Public Art

This is absolutely brilliant! Can you share a bit about how it works behind the scenes and where you expect to continue developing?

Report

4yr ago

GitButler

Maker

Thanks @rememberlenny and for sure, happy to share! Codeball is a Multi-layer Perceptron classifier neural network. It evaluates on a per PR file basis, taking in a very large number of inputs about the code and the author. Broadly speaking, they fall in three different categories — Basic, Derived and Categorical. - Basic — Metadata directly associated with the Pull Request. For example, number of lines added, removed etc. - Derived — Data that is indirectly associated with the Pull request. Examples include number of days since last change to file, file ownership ratios, file volatility score, author-file score etc. - Categorical variables — Derived categories for the Pull Request, for instance the file type categories. A more advanced example category is the perceptual hash of file diffs. We used millions of pull requests from open source projects in order to generate training data - the predictors as well as the known outcomes (whether the PR was merged without changes/objections). Having seen tons of success and failure patterns makes Codeball really precise in evaluating contributions it has never seen before. Right now Codeball can only classify (and trigger approval of) safe PRs. We want to take it further and give the developer feedback on how to improve their code. Here is a comparison with the automation level for autonomous cars ;) -> https://codeball.ai/what

Report

4yr ago

Digger.dev

This is so awesome and makes so much sense! Indeed there's a whole category of PRs that are trivial but you also don't want to break the good habit and so teams are ending up wasting lots of time

Report

4yr ago

Polar

Maker

Thanks @igorzij, and I couldn't agree more. Bringing more automation further up the toolchain is a super interesting area, products that are "shifting left" are and will continue to be on the forefront of the industry!

Report

4yr ago

GitButler

Maker

Hey Product Hunters! 🌊 As engineers, we hate being blocked and waiting for Pull Request reviews, especially when the changes are obviously safe. We thought — if an AI can drive, surely we can train one to review code! 🧠 Codeball is a neural network trained on millions of Pull Requests (“wow, big data!”). It accurately predicts PRs that would be approved without further feedback. 🧩 Use the Codeball GitHub Action and auto-approve your good PRs (https://github.com/sturdy-dev/co...)! 🎯 Optimized for precision — Codeball only approves PRs that it’s really confident in. It uses over 100 different predictors (e.g. the code diffs themselves, author experience with the changed files as well as past issues with the type of change) for each contribution it checks. 🤯 When you stop the rubber stamping fakearoo on most PRs, you start actually paying more attention to tricky ones. This avoids the phenomenon informally known as “omg we have a bug in production!”. 🎮 Play with the live demo at https://codeball.ai to see how Codeball would have performed on your repository and how much time (and money) you would have saved last month. If you like it, then you shoulda put the GitHub Action on it. 💸 Codeball is free for open source projects and costs $10 / user / month for private repositories. It has a trial, of course, and because we are ballers, Codeball only costs if it actually saves you ⏰ & 💵 (there's a pretty dashboard)! Thanks!

Report

4yr ago

Does this also work with DevOps?

Report

4yr ago

Polar

Maker

@kenny_strubel Codeball is currently in beta, and is only available for GitHub at the moment. But we have both DevOps and other platforms on the roadmap coming soon. Please reach out to us in DM if you're interested in joining the invite-only beta for DevOps. :-)

Report

4yr ago

@zegl That sounds great! Where can I shoot you a dm?

Report

4yr ago

Polar

Maker

Nice! You can DM on Twitter @zegl or email gustav@getsturdy.com

Report

4yr ago

Great application of AI. Excited to give this a spin.

Report

4yr ago

Love it! Will be sharing w our eng team asap! Always love seeing what you’re working on @videlov !

Report

4yr ago

1 2 3 4

Reviews