VibeShift
Get secure, working code in 1 shot
82 followers
Get secure, working code in 1 shot
82 followers
Put cursor in auto feedback loop so that it tries to generate secure, fully working code in one shot. Completely open source.
Free Options
Launch Team / Built With
Render Sandboxes: a computer for every agentGive every agent its own machine. Starts in under a second
Promoted
Hey Product Hunt 👋
I'm Om, an indiehacker and I love dabbling with any shiny new toys in tech. One main problem with AI coding was that AI always made simple errors that I had to painstakingly test and then copy paste the error to AI. I wanted to just test out the final product with no errors, so I built vibeshift. I thought the natural next progression after AI coder is the AI tester.
It sucks that we humans have to test and fix AI generated code slop. If we are using AI to stitch together prototypes, then the least AI could do is check its mess and give us working, secure prototypes.
I am building this so that fellow indiehackers or "vibe coders" could ship prototypes without the fear of broken or insecure features. This is completely open source product but for those who want to test remotely and test larger codebases for security issues, I have built a cloud based tester for parallel testing.
I'd love your feedback, questions, or thoughts on how to make VibeShift even better. I'm always interested to talk about anything tech. Just hit me up at X (Omiiee_Chan).
Thanks for checking out VibeShift 💙
I've been wondering why AI coding solutions always make me say 'it still doesn't work', instead of testing the code first.
Thanks for your effort. Congratulations @ilikepizza2 :)
lol and the endless copy pasting console errors.
@pritraveler Thank you for your appreciation :]
The concept of an open-source code generator that aims for secure and fully working code in a single shot by leveraging an "auto feedback loop" sounds incredibly promising for boosting development speed and reducing debugging time!
Thank you @supa_l, that is our aim, to keep up with the speed with which AI codes :)
I've tried Ollama v0.7 and really appreciate how it's geared toward secure, fully working code out of the box. The new engine feels like a true step forward—multimodal support with models like Llama 4 runs smoothly, and the fact that it’s completely open source makes it even better.
Automated feedback loops like this push coding tools closer to true self-correction and reliability from the start.