Aperture
Product Hunt Witchayut
Aperture is a new open-source AI code editor. It stages changes as diffs, then checks them before you apply anything.
Based on reporting by Product Hunt, Witchayut — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Aperture launched today on Product Hunt as an open-source AI code editor built around a very specific promise: let the model make the change, but don’t let it sneak anything past review. Its Composer feature plans edits and stages them as a diff instead of dumping loose suggestions into your project.
Before you accept the change, Aperture runs a series of checks. It parses the code, checks whether imports resolve, runs type checks, renders the preview, and then runs tests. That last part happens in your browser tab, and the company says it’s free.
That matters because AI coding tools have a bad habit of sounding confident right up until they break something subtle. Aperture is trying to make the machine-generated part feel less like a leap of faith and more like a controlled edit.
The launch page also places it in the AI coding agents and AI code editors buckets, which is probably the right neighborhood. The pitch here isn’t that the editor replaces judgment. It’s that it puts a gate in front of the usual chaos, and asks the code to prove itself first.
My take — AI-written commentary, not fact-checked reporting
Open source is doing the useful work again: not louder demos, just tighter guardrails. Aperture reads like the right answer to AI coding hype, because the real product is skepticism with a UI. That’s a healthier trend than pretending every prompt deserves to be merged.
Read more about this at: Product Hunt