I am experiencing an issue where my repository is not deploying to the main branch of my GitHub as intended. Instead, the system is automatically creating and switching to a secondary branch (e.g., gh-pages or a feature branch) and deploying there.
The specific behavior I’m seeing:
When I push code, a new branch is sometimes generated automatically and deploys to the vercel supplied domain.
The live domain is linked to the main branch and therefore my edits are never deployed.
I have verified that main is set as the default branch in my GitHub settings.
For some reason deploying to the second branch and then again to the main doesn’t work, as the changes never deploy. v0 pulls the old version off GitHub before it can deploy the new changes.
I have tried turning off the automatic push/pull to see if that would make a difference and it did not.
I have tried clearing my cache.
I have tried creating a new project on GitHub.
I am at a loss…
Request: Could you help me identify if there is a GitHub Action workflow or an Environment setting that is overriding the deployment target? I want all successful pushes to main to update the primary domain, without creating auxiliary branches. I recently switched my project over to a new account and ever since doing that, everything with git and deployment has gone crazy.
So then your second branch in gh is also connected to a duplicate chat. Then you can push all update to that branch. merge to main when ready, and then pull in merged changes to you rmain chat.
So it worked for awhile then decided no more. It insists now that I do a preview mode on a new git branch before I can put anything to production. When I try to put it to production, it does not deploy to my domain, but a temp vercel domain. I do not want this. v0 should not be creating branches on my behalf/publishing to a branch that I have not it set to publish to. Furthermore, when I go to manually try to push the changes to the main git branch and then publish it on v0 and then it goes right back to the “new” git branch that was created. Is there anyway someone can look into this issue on my account? I am paying for a team plan.
and you have checked that all the auto commit / pull settings are turned off? Have you tried unlinking the branch and then relink to a new one maybe?
Really weird honestly as I have never had this happen where when I deploy it auto creates a new Git branch. maybe its a setting you have set somewhere in Vercel specifically on that project by chance?
I know this may sound like a dumb question, but have you tried asking the v0 agent why this is happening? I mean…maybe its worth a shot? Just a thought, but honestly not sure. All so odd.
Hoping someone else from the community or Team will see this later or Monday and be able to hop in.
I actually did ask the agent for help regarding the github and it said that was outside of it’s scope and to reach out to support.
Yes, I checked that all the auto commit/pull settings are turned off. Yes, I tried unlinking the branch and then relink to a new one, I created 3 new repositories, does that count?
I have also having been issues getting my Supabase to connect, it tries to create a new account every time I’m prompted to install/connect Supabase (even though it’s already been installed) the variables are all in place in v0 and vercel. “Install Supabase” is the agents response to a LOT and it’s frustrating bc it is installed and the variables on both vercel and v0 are correct. Even on the supabase account it has that little vercel triangle indicating it’s connected.
I really hope to get to the bottom of all this as it’s become really hard to create anything without these two things in place.
I would love the opportunity to have someone to take a look at my account and try to solve these issues, as I am a paying customer.
See my attached moments of Supabase frustration below. I also made it a rule that “Install Supabase” cannot be the answer, that did nothing. I’m not sure what I am doing wrong when it comes to integrating these third-party tools.
Hmmmm ok yeah this is VERY odd for sure.
Random question, what agent are you using? I dont think the solve to this will be the answer to what you say here, but I am just curious.
As for what could be happening, again, this is very weird and odd that it keeps repeating. Even when you have disconnected and reconnected.
I would say to try just starting fresh (import your code from github), but I understand that may not be a route you want to try. More just to try for more trouble shooting.
As for having someone from the Team look into this, I fully agree with you. I think with Holidays upon us, their response time may be a bit delayed. But I am certain they will see this and respond asap (I of course understand the frustration in the waiting though).
In the meantime, I want to see if some others from the community may have some ideas. @ussaaron ,@avik-creator, @hasriza any ideas here?
I use the v0 pro agent, I’ve tried the v0 max and it didn’t seem to make a difference. It just doesn’t seem to listen to me anymore, in the sense I ask it to do a very simple task and I’m ignored/nothing happens. I also keep getting “page unresponsive” when I’m on the v0 page for editing and the AI is thinking. I really am at a loss, but I cannot keep purchasing credits if it’s not going to operate.
I did recently transfer this project over from one account to another on v0. I did this so my v0/vercel accounts would both be on the same accounts, but should/would that have cause an issue? They should be on the same team accounts right? This is the only thing that I can think of that has changed.
There is clearly a bigger issue going on here and I don’t think I’m going to be able to solve it on my own, so I appreciate your response and for reaching out to your team members.
However, I am experiencing a recurring issue with the v0 GitHub integration where the tool is preventing me from deploying directly to my main branch.
Even when I have main set as my active branch in both the v0 and GitHub, attempting to ‘Publish’ or ‘Push’ triggers a forced redirect to a ‘Preview Environment.’ This automatically creates a new feature branch (e.g., v0-all-changes) and switches my active working branch to this new one.
If I manually switch back to main, v0 insists on a test environment deployment again or re-deployment to the production environment, which automatically puts git back to the “newly created” branch and traps me in a loop where it deploys to a “Preview URL” vs my “Production URL.” I would like to restore the ability to push/deploy directly from my primary branch to production as I was able to do previously. Is there a project-level setting or a way to disable this forced branching behavior?
@hutchisonmallory-557 So you have selected “Production” and it’s still deploying to preview? That sounds like a bug to me, hopefully the team can help you resolve this!
(You’ve selected “Production” in the publish menu?)
That is correct. I select “Production” and it deploys to a Preview URL even though I have a Production URL set to my account.
However, It will deploy to production if it is not Synced to GitHub, but as soon as the sync is active, the loop begins. I am using Vercel under ‘Project Settings’ to initiate the initial connection/sync with my Git account, and it appears correctly in my v0 dashboard shortly after, so the connection itself seems valid.
The branching issues have made deployment very hard. Even more concerning, the AI isn’t answering to simple tasks, I’m getting ‘Page Non-Responsive’ errors when trying to run queries, and I’m told by the AI (on repeat) that my Supabase integration is not installed, even though it has been fully integrated. I’m wondering if this is a larger infrastructure issue.
If you all could look into this, I would truly appreciate it!
Yeah.
I cant replicate this issue at all. I have actually tried to haha. And all my deploys work fine and dont create new gh branches.
Not sure what is going on tbh.
I’m happy to keep building in v0 without Git for now and will export the code manually until things are settled. My main priority has shifted to getting supabase connected so I can keep building. I’ve been running the logic manually in the Supabase dashboard, but to get the level of functionality we’re after, I need to link the environments directly. As you can see below, my Supabase project is already present; however, when selected, I get the red error message.
I’ve successfully connected Supabase and moved away from Git. However, I’ve hit a deployment ‘split-brain’ issue: while my frontend changes go live instantly when I hit publish, the backend/API logic is not deploying.
I have tried clearing the cache and redeploying on my own, and it has done nothing. I’ve also verified the live setup by changing a visual on the main page - that update works, but the backend seems to remain stuck on stale code (see snips below. Please note: the edits it made to try to fix this issue did not work).
Since I am no longer using Git to trigger builds, what is the best way to force a full rebuild of the backend so the server-side logic matches the frontend since the deploy button doesn’t work? Also, has anyone on your team been able to look further into the ongoing Git issue? It would be great to return to a standard workflow!
hmmm ok this is definitely a head scratcher.
I am not sure what is going on here to be honest nor the best way forward.
I want to try asking two others. @ussaaron from the community I feel may have an idea and / or @jacobparis from the Team may have a solid solution right away or be able to help move this to the right person.
If we dont hear anything in a day or two on this, I will reach out to the Team directly for you.
Sorry for delays on this and understandable frustrations.
I could probably solve this pretty quick if I could look at the actual code, but without that I’m only guessing. If there are sync issues between your v0 project and github or Supabase there are a few troubleshooting steps to try first. First check if you can execute sql actions directly through the sql editor in the supabass dashboard, you want to rule out RLS rules that may be triggering silent errors. Next you want to rule out v0 fork issues so duplicate your chat, and try again from v0. If that fails the next thing I would do is start a whole new v0 project and try to recreate your current environment.
Also you can check for logging errors by opening console log and going to the network tab and checking the response body of your requests. I’ve debugged a lot of v0 issues this way because the response errors are usually more verbose when accessed through the network tab.
