TLDR; Ellipsis is a GitHub app that automatically reviews, summarizes, and answers questions about pull requests.<p>Hi HN, @hunterbrooks and @nbrad here, cofounders at Ellipsis. We’re on a mission to build an AI software engineer. So far, we’ve built the automatic PR review functionality we’re sharing with you today.<p>Whenever a PR is opened, Ellipsis will automatically:<p><pre><code> - append a summary of changes to the PR body,
- perform a code review, checking for best practices like DRY, descriptive variable names, etc.
- check for violations of any of the custom rules you’ve provided
</code></pre>
The most common question we get is “does it work?”, so we’ve added a 7 day free trial. Here are some examples from popular open source repositories within the past week:<p>- <a href="https://github.com/hyperdxio/hyperdx/pull/326">https://github.com/hyperdxio/hyperdx/pull/326</a><p>- <a href="https://github.com/relari-ai/continuous-eval/pull/43">https://github.com/relari-ai/continuous-eval/pull/43</a><p>- <a href="https://github.com/jxnl/instructor/pull/467">https://github.com/jxnl/instructor/pull/467</a><p>- <a href="https://github.com/ion-design/numi/pull/1">https://github.com/ion-design/numi/pull/1</a><p>You can also tag @ellipsis-dev in a comment and ask it for a re-review or to answer a question about the PR. You’ll get a new review for every commit that’s pushed to the PR branch. Our docs (<a href="https://docs.ellipsis.dev/review">https://docs.ellipsis.dev/review</a>) have more info.<p>Today, 17 companies, including PromptLayer and Warp (W23) are using Ellipsis. We’re working on the code generation component (ex: tag Ellipsis to fix bugs), but that’s still in a public beta.<p>Ellipsis doesn’t store or train models on your code.<p>Ask: Try it out, it takes 2 clicks to install at <a href="https://www.ellipsis.dev">https://www.ellipsis.dev</a>. Feedback? team@ellipsis.dev
We use Ellipsis across most of our repos. PR titling and summaries are great and definitely reduce the cognitive load for both code contributors and reviewers. Reviews are also a good first line of defense for code quality, catching issues human reviewers haven't caught. Though on occasion it's also a little overly pedantic or confused. I guess not unlike us humans at times ;-)
This looks great.<p>I've been getting a lot of ads for a product with a similar premise ("AI-first Code Reviewer"): CodeRabbit.ai. Can you help me understand how this product compares?
I guess it's clearly within the capabilities of modern LLMs, but it's still cool to see that you can just specify linter rules with natural language now.