TL;DR: Review is an open-source desktop app designed to give you confidence in your code. Critique interactive visualizations, diffs, and agent traces, all from one interface. It’s like Cursor, but purpose-built for code review.

Our Ask: Try our app on your toughest code reviews; it is most illuminating for large, complicated changes. Download the app, star the repo, & ask your questions on our Discord!

Demo Video: https://youtu.be/mDgxhzrFi2Y

The Problem:

No one knows how much code we should be reading these days.

It feels tedious and old-fashioned to manually comb through thousands of lines of changes, and yet irresponsible to merge code you can’t confidently describe.

We’ve lost touch with how our own systems work.

Building a company is a long game. Development is an iterative process. It takes accountability and code comprehension at every stage to build a great product that customers can count on (ex 1, ex 2).

Our Solution:

Before AI, staff engineers understood massive software systems even when they didn’t have time to read all contributed code. They prioritized without sacrificing complexity by:

- Scrutinizing RFCs and architecture diagrams

- Skipping low-impact PRs

Review treats you like a staff engineer. You can think of it as a ‘Cursor for code review.’ It helps you build programmatic, customizable RFCs from your PRs using coding agents.

Instead of reviewing 1000+ line diffs you:

1. Consult a variety of interactive diagrams built on our SDK:

2. Chat with the authoring agent to leave comments on the change:

3. Audit the agent’s decisions by connecting traces to code:

—

Team:

Sid (CEO) was a tech lead on Palantir’s cloud infrastructure team, where he built secure hybrid cloud deployment infrastructure. He studied CS at Harvard.

Milan (CTO) was a cofounder of Bolto (YC S23) and previously worked on Palantir's ML infrastructure team, where he built tools for in-platform model deployment and evaluation. He studied CS at Harvard.

Ketan (CSO) has worked as a machine learning engineer at Robust Intelligence where he red-teamed AI systems to make them fail, and at Snowflake where he built data-querying agents. He studied CS at Stanford.

Alex (CPO) worked as a software engineer at Pinterest and at Flowcode where he led an effort to rearchitect the product for large enterprises. He studied CS at Harvard.

—

We’re going all in on open-source. Our dream is to be the open-source code forge for the AI era.

Install the app here: https://install.dev.fast

Have questions, comments, or interested in a design partnership for our hosted product? We’d love to hear from you!