Inspiration
Code review tells us what changed, but it rarely proves the feature actually works. I wanted a way to test PRs autonomously
What it does
It tests all the PR changes in a snadbox environmnet with all the tools the agent needs and posts the testing videos on the PR
How we built it
We built it using codex and gpt-5.6 models. i started with brainstorming this idea, because no one is solving this. Grilled on the premise and then did a bunch of iterations before landing on the design.
Challenges we ran into
We ran into chrome browser challenges while providing agent tools. then preview url access to agent
Accomplishments that we're proud of
I am proud of I was able to ship this in very little time
What we learned
I learned that with AI the bottleneck is to understand what to build
What's next for E2E
I would want to add more sandbox providers like daytona, sprites.dev etc and provide this to more people for testing and feedback
Built With
- node.js
- playwright
- react
- typescript
- vite
Log in or sign up for Devpost to join the conversation.