I am encountering an unusual problem. while working in Firestudio today I began to encounter a 404 error message when my app loaded. The ai tried multiple times to correct this problem but was unsuccessful. I then decided to revert back to a previous working commit. After that the 404 error was still occurring. I restarted the virtual machine, tried reverting the commit again but I am stuck with the 404 error.
Hi @PPars ,
This seems like an issue in code rather than an issue with Firebase Studio or a gemini model.
404 means “Page Not Found”. In other words, the link isn’t available. Are you working in a NextJS project?
According to may understanding of the development environment, when the checkpoint is reverted then all of the code is reverted as well. I have reverted the software to a previous working commit but the issue remains.
can you share a screenshot?
Thanks, I was able to resolve the error to continue working on the project. However, I do not know what the actual problem was (or is)
. I imported the project repo into a new project and it works in the new project space.
Yeah I do not know the error you encountered either. But glad it started working for you!
I imported the git repository into another project and it worked fine there. Eventually I decided to try cleaning out the git repo with the following commands and now my original project is working again.
git clean -fdx
npm install
When a Firebase Studio project fails to revert to a checkpoint correctly, it’s usually because either:
-
Untracked local changes are blocking the rollback, or
-
The checkpoint metadata itself wasn’t applied fully (e.g., mismatch between Firestore + Hosting states).
Steps to try:
-
Ensure Clean State → Run
git status(if connected to GitHub) orfirebase status(if CLI linked) to check for pending changes. Stash/commit them before reverting. -
Force Checkpoint Reset → Many users forget you can use:
firebase experiments:reset --checkpoint <ID>
(replace <ID> with your checkpoint hash).
3. Verify Target Environment → Make sure you’re reverting to the correct env (dev, staging, prod). Reverts only apply cleanly if the env matches.
4. Clear Local Cache → Delete .firebase/ or .cache/ dirs, then retry.
5. Last Resort → Manually restore config from checkpoint commit (if GitHub-backed), then redeploy.
Common Pitfall: Gemini and Firebase Studio sometimes only roll back files, not linked services (Auth, Firestore rules, Functions). After a revert, always run:
firebase deploy --only firestore,functions,hosting
to ensure parity.
If it keeps failing, drop the specific checkpoint hash in our Signal Beacon Slack channel — we can help walk through recovery.