AI Studio keeps saying "The service is temporarily unavailable. Please try again." and "There was an issue while sharing your applet"

I can’t work anymore as its continues indefinitely to say “The service is temporarily unavailable. Please try again” whatever I do. Even when i just ask a question!

My published version still works but I can’t rollback as it also fails to restore from the last versions!

‘git status’ gives:

fatal: not a git repository (or any of the parent directories): .git

Some errors I see in the conversation (action history):

  • “Failed: checking task status”
  • “Failed: viewing entire MCP router file”
  • “Fixing the Dev Server: I’m focused on why the development server failed to start, keeping in mind the user’s previous query about resolveMcpUserFromRequest. I’m going to investigate the recent edit to server.ts to see if that’s a key factor. The priority is to debug the startup issue first, and then circle back to the function call.”

The Preview is down but the code is visible (downloaded).

I’m completely stucked!

I’ve tried to hard refresh, to use a private session, to rollback manually to old code but nothing works!

Do I need to start a new chat window? To click on Remix? To create a new project?

Google AI Studio never did that before! Is this message “The service is temporarily unavailable. Please try again” caused from a Google server problem? Or is it on my side? I’ve noticed that this Google Developer forum is failing sometimes tonight (not accessible). My other projects work though.

Please help.

same issue here : The service is temporarily unavailable. Please try again

I’m also getting the same The service is temporarily unavailable. Please try again, at this time even though the cloud report says the errors were resolved.

The problem has been solved on my side thanks to the intervention of @Mahesh_Sutar :clap:

It seems that he restarted the DEV sandbox which has stopped on itself provoking this error of unavailability. Give him your APP ID and screenshots, he will solve your problem.

I’m also getting this, wondering if someone can help me out at all?

Hello @Mahesh_Sutar,

I’m experiencing exactly the same issue reported by Larry Baxter.

My Google AI Studio applet repeatedly displays “The service is temporarily unavailable. Please try again”, along with “There was an issue while sharing your applet”.

Other apps under my account work normally. I also created a Remix, but the issue persists there.

Affected Applet ID: 5e8c632e-9080-4de0-bc93-71c1277cfb52

I have already submitted two feedback reports through Google AI Studio.

Larry mentioned that his issue was resolved after a DEV sandbox restart.

Could your team investigate whether my applet has a similar sandbox or workspace issue and restore its functionality without modifying or resetting the project files, version history, Firebase resources, or publication?

Thank you.

Hello @Edixon_Rodriguez ,

Could you please try accessing and saving your applet (5e8c632e-9080-…) again now, and let us know if the issue still persists?

I’m experiencing the same Google AI Studio issue with my Hoogle application.

Applet ID: 9f536731-531e-484f-a80f-e3629e5d5c..

I’m repeatedly receiving “There was an issue while sharing your applet.”

The development environment also fails to provision with the following error:

Failed to save env vars: open .dev.env.json: no such file or directory

My production application is currently working on Cloud Run, but AI Studio’s development environment and sharing functionality are failing.

I’ve attempted refreshing the workspace without resolving the issue.

Could the Google AI Studio team investigate and restart or repair the development sandbox associated with this applet without modifying or deleting my project files, revision history, environment configuration, or existing deployment?

I would appreciate assistance restoring the original project rather than having to recreate it.

Hello @Hoover_Electric ,

Can you share screenshot of error ?

Hello @Mahesh_Sutar,

Thank you for your response on October 14.

Unfortunately, my SEVEN EMPRÉSTIMO SUPER applet is still experiencing serious problems. I would like to provide a more complete timeline because the situation has become increasingly disruptive.

Affected Applet ID: 5e8c632e-9080-4de0-bc93-71c1277cfb52

1. Intermittent recovery followed by recurring failures

At one point, the AI Studio development environment appeared to recover and worked for approximately 1–2 days.

However, the problems returned, and the failures have since become more frequent and severe.

This suggests an intermittent issue rather than a permanently inaccessible project.

2. Current AI Studio errors

The applet repeatedly experiences:

  • “The service is temporarily unavailable. Please try again.”
  • “There was an issue while sharing your applet.”
  • “An internal error occurred”, sometimes after 0 seconds.
  • “There was an unexpected error. Finish what you were doing.”
  • “1 error running the code.”

Gemini sometimes fails before completing even a simple read-only inspection.

3. Workspace and file operations

Problems also occur when attempting to:

  • Read or inspect source code.
  • Upload or work with Markdown (.md) documents.
  • Modify application files.
  • Delete files directly through the code editor.
  • Run normal development tasks.

These failures have made the applet extremely difficult to maintain.

4. Remix did not solve the problem

I also tried duplicating/remixing the applet, but similar errors appeared in the duplicated project.

This makes me concerned that the problem may involve project state, workspace metadata, dependencies, or the development sandbox.

I cannot determine the root cause from my side.

5. Possible PRE synchronization problem

There is also an unresolved synchronization concern.

The JavaScript bundle observed in the PRE environment differed from the locally generated bundle during a previous investigation.

A code change and diagnostic marker present in the workspace were not observed during subsequent manual testing in PRE.

Therefore, I cannot confirm that PRE is receiving the latest workspace changes.

PRE URL: https://ais-pre-7tps7sfqyq336kcrvfn5qb-411029201667.us-east1.run.app/

6. Other applications work normally

Other AI Studio applications under the same Google account continue working normally.

I have submitted approximately six support/feedback reports over the past two weeks, but this particular project remains blocked.

Request for engineering investigation

I noticed that another developer, Larry Baxter, reported a successful recovery after an apparent DEV sandbox restart.

Could your engineering team investigate:

  1. Whether this applet’s DEV sandbox needs to be recovered or restarted.
  2. Whether workspace metadata, filesystem access, or internal locks are corrupted or blocked.
  3. Why Gemini fails immediately during read-only and file-editing operations.
  4. Why applet sharing/build operations repeatedly fail.
  5. Whether the latest build artifacts are correctly reaching the PRE environment.

Please prioritize a non-destructive recovery.

It is essential to preserve all existing source files, version history, Firebase resources, configurations, operational data, and production deployment.

If possible, please let me know whether this issue has been escalated to the relevant engineering team and what recovery options are available.

I am close to completing the MVP, but these persistent platform failures have blocked development for approximately two weeks.

Thank you for your assistance.

(post deleted by author)

Hello @Edixon_Rodriguez

If your project contains any files or folders with a leading slash (/) in their path (for example, /docs/... instead of docs/...), please remove or rename those files in the File explorer so their paths do not start with /, or restore an earlier version from the Versions tab in the left sidebar and re-add the files using relative paths without a leading /.

“”"

Hello @Mahesh_Sutar,

Thank you for your guidance.

Unfortunately, my SEVEN EMPRÉSTIMO SUPER applet is still completely blocked.

I checked the File Explorer and previously exported project ZIPs. No leading-slash paths were identified in the exported files, although I understand that internal workspace metadata may differ.

I also tested the project again after your response:

  • Gemini 3.8 Flash fails with “An internal error occurred” after 0 seconds.
  • Gemini 4 Argon fails with the same error after 0 seconds.
  • “There was an issue while sharing your applet” persists.
  • Even read-only file inspection cannot complete.

I therefore cannot reliably edit, rename, remove, or save project files.

Could you please check whether invalid paths exist in the applet’s internal workspace metadata and investigate the DEV sandbox directly?

If no invalid paths are found, could this issue be escalated for a non-destructive sandbox recovery?

Before attempting a version restore, I need to understand how to avoid losing recent project changes.

Applet ID: 5e8c632e-9080-4de0-bc93-71c1277cfb52

I am still unable to continue developing my MVP.

Thank you.

Hello @Mahesh_Sutar,

please I need your help. My Google AI Studio project is showing “There was an issue while sharing your applet” and “Invalid argument.” My project is still visible under My Apps, but it will not open. I saw that you helped other developers by restarting their development sandbox.

My App ID: 53268c2c-8224-4be2-a078-02b6612dbbfa

Project URL: https://aistudio.google.com/apps/53268c2c-8224-4be2-a078-02b6612dbbfa?showPreview=true

Please help me investigate whether my development sandbox needs restarting. I want to recover my existing project without losing my source code or data. Thank you very much.

Hello @Precious_Effiong ,

I have DM`ed you for more details .

Hello,

@Edixon_Rodriguez, your app (5e8c632e-908...) may still contain files with a path that starts with / even if they don’t show up in your exported ZIP, for example Markdown documents you added to a docs folder. In the Versions tab in the left sidebar, restore the last version from before those documents were added, and re-add them with relative paths such as docs/... (not /docs/...). Restoring a version first saves your current work as a version too, so your recent changes are not lost and you can switch back to them.

@Precious_Effiong, the preview of your app (53268c2c-822...) may fail if one of your components has a syntax error, such as an extra or missing bracket. Please try restoring the last working version from the Versions tab in the left sidebar, then click “Start new chat” and try again. If it still fails, please let us know the approximate time it happened.

@Hoover_Electric, are you still seeing the error on your app (9f536731-531...)? If so, could you please share a screenshot of it?

Hello @Mahesh_Sutar, i have the same issue aswell, it says there was an issue while sharing your applet. Here’s the applet id a36fac5-d52d-4fc2-9397-03d2ae9cef5f

please fix this as soon as possible.

Thankyou

Hello @Mahesh_Sutar,

I have now followed up on your recovery recommendations and performed additional diagnostics.

The applet is still not usable for normal development.

Here are the new findings:

  1. I manually deleted an old Markdown file. The deletion was successfully saved and persisted, but Gemini continued failing.
  2. I attempted restoring an earlier version through the Versions tab. The version history and file state changed, but the agent did not recover.
  3. I inspected exported ZIP files and found no leading-slash paths. I understand that internal workspace metadata may differ.
  4. Gemini 3.8 Flash successfully ran a grep command, returned valid results from App.tsx and test files, and then failed with “An internal error occurred”.
  5. Other attempts still fail immediately, sometimes after 0 seconds.
  6. The PRE application is accessible and its Route Admin navigation now works, but the AI Studio development agent remains unreliable.

Applet ID: 5e8c632e-9080-4de0-bc93-71c1277cfb52

At this point, could your engineering team please inspect the internal task execution logs and workspace metadata, and attempt a non-destructive DEV sandbox repair or restart?

If this cannot be resolved directly, please advise on an official recovery or escalation path that preserves the current source code, version history, Firebase resources, and production deployment.

I have been unable to continue normal MVP development for over two weeks.

Thank you for your help.