✓ I can execute sql actions directly through the sql editor in my supabass dashboard.
I have tried forking; however, when I try to fork, the SQL scripts refuse to run. I believe this is due to the CSP error below. Gemini seems to think the issues lies in my middleware file. However, when I manually made the suggested Gemini edits, and requested the AI do it, nothing changed.
For some reason when I fork, a new v0 project is created… Not sure what I’m doing wrong there?
If @ussaaron would be able to connect or has any recommendations specific to the v0 console errors below, I would appreciate it!
@hutchisonmallory-557
I have reached out to the team now regarding this issue. Hoping they will be able to jump in with some answers and solutions for you here. Sorry for any inconvenience you are experiencing due to this issue.
This issue continues to occur. Although I’m beginning to wonder if this is not related to my code, as I the same issues show up when I have trouble deleting my forks, which now show up as projects on my account.
Actually, I’m just now realizing that my project on v0 is gone. They are all forks now, I’m not sure where went/when it disappeared. Also, when I go to supabase, and go to the dropdown they are all forks listed. Shouldn’t they be the projects?? I could be wrong here! Perhaps this is why it doesn’t want to connect?
Vercel is displaying all my forks as new projects as well.
Perhaps I do not understand what a fork is. I would like to clarify that the “New Chat” is what we call a fork and it should not create new projects on vercel/v0 each time it is selected?
Hmmm interesting. I have never used the “new chat” in the version history as a way to fork. I have always just duplicated (what used to be called fork in v0), and thus have never had this issue. but if youre using “new chat” from the version / chat history, I wonder if this is the culprit behind it all. It seems like we may be getting closer to the problem at had. Lets see what they team comes back with and / or if you can try another method (duplicate rather than new chat) to fork and see if that works. Just more trouble shooting I think, but I feel we are getting there.
Could I verify where the following API keys are located on supabase? I’m currently using the same two keys to verify the following 4 keys. I’m wondering if this could be the issue? Also, I have switched to “duplicating” vs “new chat.” This hasn’t helped with the supabase issue so far, as I still get the SCP and API errors in my console log, but I will stick to this proper method in hopes that it might clean up the process.
NEXT_PUBLIC_SUPABASE_ANON_KEY
SUPABASE_ANON_KEY
NEXT_PUBLIC_SUPABASE_SERVICE_ROLE_KEY
SUPABASE_SERVICE_ROLE_KEY
The following keys are on supabase are being used for all four areas, as I’m unsure what is supposed to be for the NEXT_PUBLIC_SUPABASE_SERVICE_ROLE_KEY and NEXT_PUBLIC_SUPABASE_ANON_KEY.
This past month all credits have gone towards account diagnostic/feedback (I have spent $20 today and am already being prompted to to purchase more).
Unfortunately, I like to eat and cannot afford this.
While we were able to get git and supabase connected, it still appears that v0 cannot execute SQL commands on supabase and v0’s AI still struggles to complete very simple tasks.
AI 5 edits in: “I’ll add a debugging tool…. (even though one was already added the second edit in)” Like no, fix the problem dude lol.
I will most likely be switching to a new product by the next billing period. Do you all have any recommendations as fellow developers?
Hi @hutchisonmallory-557 , really sorry to hear this. Of course this is not what we in the community or team want to happen. I fully understand your frustration and feelings on this. I’m going ping the Team right now and see what can be done to solve this and maybe more.
I admit we are getting places, so thanks everyone for your help and for looking into this.
Here is where we are at:
Supabase is now connected.
Git is now connected.
v0 is able to publish without creating a new branch each time.
v0 won’t let me “Run Files.”
Every time I try to fork (even doing it the correct way) the AI starts acting funny/messing up my work. I always have to go back to my old chat.
v0 is unable to use AI efficiently. It’s very hard to complete a simple task and takes many prompts. That said, there are almost 800 prompts on this one chat as I cannot fork/run the files.
v0 is unable to execute commands to supabase and lies that is does.
It is my understanding that v0 is over here trying to create RLS polices on supabase and is indeed charging for it; however, to my knowledge this is not possible. AI has instructed me to go try to click all 40 scripts and go run them manually myself, but when I “click” them, they won’t open. Also, that sounds like a real pain lol.
I’m just wondering what is going on, on your end???
Supabase is now connected.
Git is now connected.
v0 is able to publish without creating a new branch each time.
Those “Run Files” appear to be gone.
I’m able to fork now using the new git functionality.
Buuuuttttttt
I cannot get my v0 page to load for the life of me. It is “Page Unresponsive” or just a blank page. I have cleared my cache/browser history. I have upgraded to 1g fiber internet. I have tried a new browser. It only happens on the v0 page.
At one point I did get v0 to briefly load; however, the AI wasn’t displaying the edits it was saying it was making.
I have tried submitting a support ticket via v0 but the AI doesn’t give you an actual “submit” button after the view case form. If you you all could escalate this info, it would be appreciated.
I’m following up because the platform’s performance isn’t just unstable—it’s regressing. Rather than improving with updates, my experience is getting worse, and I am now completely blocked:
Access Blocked: I am currently getting an “unauthorized” error when trying to create a new branch from my projects page, despite being logged into the correct profile. I can no longer access the editor at all.
AI Logic Failure (Ghost Edits): When I recently had access to the chat, the AI became a liability. It would claim to have applied fixes, but no changes were reflected. Worse, when it did make changes, they were often incorrect or broke existing functionality, making the application worse instead of better.
I’m a huge fan of v0’s potential, but it has reached a state where it is actively hindering my progress rather than helping.
Could you please look into why my account is hitting this “unauthorized” wall? I’m happy to provide my Project ID to help resolve this.
Update to my last post: The ‘Access Blocked’ issue has resolved itself, which is great.
However, I’m still hitting the AI Logic Failure (Ghost Edits) which has broken some existing functionality. I’m trying to roll back to a version from before the Git changes were implemented to start fresh, but I’m stuck—I can publish the old version live on Vercel, but I can’t seem to get that version back into the v0 editor.
Any advice on how to force v0 to ‘revert’ to a specific previous deployment/state?
Currently, the ‘healthy’ version of my project is live on both GitHub and Vercel, but the v0 editor has diverged significantly. Over the last 190+ edits, the AI has essentially been ‘fighting itself’—consuming my credits while introducing Ghost Edits that have broken existing functionality.
The Issue: v0 has locked the old chat that I want to fork or branch from because it’s a “merged” branch. It literally prevents new messages from being sent to protect the integrity of the main code.
I am actually dumbfounded by the behavior I’m seeing: if the code is deployed on Vercel/main, shouldn’t that be the version reflected in the Git main branch? When I attempt to create a new project using that same Git/main source, it still inexplicably pulls the broken state instead of the actual code in the main repo.
I’ve tried to have the AI help me redeploy prior edits as a last resort, but it seems unable to ‘reach back’ and pull the clean code from my Vercel deployment into the current editor state.
The AI itself has admitted to this UI limitation:
'Unfortunately, v0’s GitHub UI doesn’t seem to have an obvious branch selector dropdown visible in this view… it doesn’t appear to be clickable.’ and 'There’s no obvious in-app way to switch brancheslike you’d expect’
The Goal: I need to force the v0 editor to sync back to the state of my last successful Vercel deployment or my GitHub main branch from my functional baseline.
Since the correct code exists externally, is there a way for your team to manually ‘re-index’ or ‘reset’ the v0 project state to match the Git/Vercel? I’d like to stop burning credits trying to fix ‘ghosts’ and just get back to a functional baseline. Two Questions:
Is there a formal v0 support line that I am unaware of for issues involving credit loss due to tool failure?
Can your team manually force the v0 editor to ‘reset’ or re-sync to match a Deployment ID?
I need to get back to a functional baseline without burning more credits on a UI that won’t let me roll back to my basline.
Hi @anshumanb, @heystu, @pawlean, @ussaaron, and @bestcodes—I know Mercury is in retrograde, which is usually my go-to explanation for everything from IT glitches to relationship issues, but I’m hitting a wall that’s becoming a real blocker for my work.
I’m struggling to roll back to a previously synced branch. Despite having a healthy, functional version live on Vercel/GitHub, my current v0 project has diverged significantly. I’m caught in a forward-only loop of edits that are breaking existing functionality and burning through my team credits.
To make matters worse, the Branch/Version selector widget seems to have vanished from my UI, so I can’t steer the project back to the correct state. As a paying team member, I’m finding it difficult to reconcile this forward-only behavior with the workflow I need to maintain my project.
Could someone please point me toward a solution? I really want to stick with you all, but I’m currently stuck fighting an editor that won’t let me revert or sync with what is currently published. What am I missing with this new Git update? Any direct assistance would be huge.