V0 is not usable anymore, it’s just producing bugs… And the worst thing is that it is not even listening to the prompts.
When told to delete something, it just add debug. When told to update something, it ruins completely other part of the app that has no connection to the prompt.
Vercel / V0 support is just ignoring any requests for help.
Last few days I spent over $300 to try to fix a simple parser bug but V0 just won’t listen. I have already invested over $2000 to the app and it looks like I should just abandon it because V0 got so much worse.
I’m right there with you. I’m having the exact same issues. I’m honestly about to give up on the new V0—the preview is terrible and constantly hits me with connection errors. I get what they were trying to do, but the execution is just a mess
From my experience, v0 work best with clear and structured prompts.
Giving too many tasks at once can lead to confusion , and without defined steps, results may drift or affect unrelated parts. Breaking things into smaller, focused instructions usually leads to more stable outcomes .
Keeping prompts simple and scoped has made a big difference for me .
That’s what I’m trying to do most of the time… simple and specific commands.. and in case of something a bit more complex I always start in the planning mode and refine the plan before starting to implement it… and it used to work well but the last update just somehow broke it.
Yes DK, if the same workflow used to work and now doesn’t, that’s definitely something on the update side. there’s likely been a shift in context management or execution boundaries. Definitely worth investigating on the system side. Pauline P. Narvas Good that the team is looking into it
Btw. the new diagnostic approach is useless. V0 doesn’t identify almost any issues. The auto-fix approach that was in the previous version was far more superior. I’m wondering why they changed it… was it to reduce cost?
I hear you, DK. It sounds like the new diagnostic approach is missing the mark compared to the previous version’s auto-fix. Hopefully, the team is taking this feedback into account, they might refine it or bring back some of the previous functionality soon.
I’m seriously about to give up on this. Look, I’m grateful for what this platform helped me build—I launched a successful project using v0—but now it can’t even handle the basics. It’s a total regression。It’s stuck in this infinite loop. It claims there’s a caching issue, says it’s fixed, but then the code fails to run. Then it apologizes, says it still sees the old file, and tries to fix it again. It’s just spinning its wheels and wasting my time。I really wanted to stick with it, but I’ve reached my breaking point. Since the updates after February, it’s been one loop after another. It keeps hallucinating progress while delivering broken code. I’m honestly done
Yes, I completely understand. I even explicitly told V0 not to mention cache issues anymore. What I’m seeing now is more of a regression in quality rather than any real progress. Over the last 3–4 days, I’ve spent over $400 just fixing issues that V0 itself introduced by modifying completely unrelated parts of the project while trying to fix something else. And yeah… Vercel won’t refund anything. At times it honestly feels like V0 is doing this on purpose just to burn more credits. I’m seriously considering moving to Claude + Bedrock.
v0 work best with clear and structured prompts.
, and without defined steps, results may drift or affect unrelated parts. Breaking things into smaller, focused instructions 
usually leads to more stable outcomes 
.
.
