Key takeaways
- The Play Console work has lead times measured in days and weeks; start it while the agent is still writing features.
- For new personal accounts, the closed test sets the minimum timeline: 12 or more testers, 14 consecutive days, then an application.
- While the test runs, Gemini CLI can handle the target API 36 migration, R8 rules, account deletion and release notes.
- Upload the first bundle by hand; afterwards the Play Developer API and a service account can automate testing-track uploads.
The Play Console work list
A terminal agent works on files and commands. Almost nothing below is a file or a command – it is forms, verification and people. That is why these items need a plan of their own:
| Task | Typical lead time | Notes |
|---|---|---|
| Developer account and identity verification | Days | Organization accounts also need a D-U-N-S Number |
| Store listing | Hours | Short description (80 characters), full description, 512 × 512 icon, 1024 × 500 feature graphic, screenshots |
| Privacy policy page | Hours | Public URL that names the app and matches Data safety |
| App content declarations | Hours plus research | App access, ads, content rating, target audience, Data safety |
| Closed test (new personal accounts) | At least 14 days | 12+ testers who stay opted in the whole time |
| Production access and first review | Days, varies | Google reviews the application, then the release |
Google's testing requirements for new personal accounts describe the closed test rules in full.
An example timeline with the agent working in parallel
Every project is different, and Google's review times vary, but a plan for a new personal account could look like this:
- Before the code is done: open and verify the account, create the upload key, write the privacy policy, draft the listing.
- First usable build: upload it to the internal track, test the signed release on real devices, fix the obvious.
- Same week: start the closed test with your 12+ testers – early, because the 14 days only count once they are opted in.
- During the test: ship updates to the closed track as Gemini CLI fixes feedback; complete all App content declarations.
- After day 14: apply for production access with concrete answers about the test, then roll out in stages.
If your organization account is not bound by the closed test, the same order still works – the test simply becomes optional instead of mandatory.
Talk to an AI Code Engineer about shipping your app to Google Play.
In a free call we look at what you built, tell you honestly what stands between it and an approved release – and whether you need us at all.
What Gemini CLI does well in the meantime
The waiting periods are a good time for the tasks a terminal agent handles best – well-defined, testable and inside the repository:
- Target API level 36. Raise
compileSdkandtargetSdk, then work through behaviour changes such as edge-to-edge layouts with tests and device checks (target API guide). - Release build health. Fix R8 keep rules so the minified build behaves like the debug build.
- Account deletion. An in-app flow, plus the small web page Google requires for deletion requests (account deletion playbook).
- Billing. Google Play Billing for digital goods, on a current library version.
- Data inventory. A list of every SDK and the data it sends, as input for your Data safety answers.
Context files and headless checks for an Android repo
Put the release constraints into the project's GEMINI.md: the permanent applicationId, SDK levels, the Gradle tasks that must pass, and a ban on reading or editing signing files. If your team moved to Antigravity CLI after Google's June 2026 change for subscriber and free accounts, keep the same rules in a file that tool reads – AGENTS.md travels best between agents.
Gemini CLI's headless mode (gemini -p) and Google's run-gemini-cli GitHub Action fit around the release: a reviewer comment whenever a pull request adds a permission or SDK, or a draft of localized "What's new" texts for each release. The build and upload steps themselves stay plain Gradle, so they behave the same no matter which agent wrote the code.
Signing and uploads through the Play Developer API
You create the upload key and keep it out of the repository; Google holds the app signing key under Play App Signing. The very first bundle usually goes up by hand in the Play Console, which also ties the package name to your app. After that, a service account with release permissions lets CI push bundles to the internal track through the Google Play Developer API – via fastlane supply or the Gradle Play Publisher plugin.
Gemini CLI can write all of that configuration. It should never hold the keystore, its passwords or the service account JSON; those belong in your CI secrets. Our article on mobile CI with coding agents shows a complete workflow.
Applying for production access
After the 14 days, Google asks how you recruited testers, what feedback you received and what you changed. Specific answers work better than general ones. Gemini CLI can help you assemble the evidence – a list of releases to the closed track during the test and the fixes in each, taken from git history – but the answers should be yours. If the application is turned down, our guide to the 12-tester requirement explains the next steps.
If you would rather hand over the Play Console part, our AI Code Engineers manage signing, pipelines, listing, declarations and the review in your own account – see the Android submission service. For iOS, read Gemini CLI app to the App Store.
Gemini CLI to Google Play: planning checklist
- Developer account opened and verified before the code is finished.
- Upload key created by a person and stored outside the repository.
- Privacy policy, listing texts and graphics drafted.
- Signed release build tested from the internal track.
- Closed test started early with 12 or more testers (new personal accounts).
- targetSdk 36, account deletion and Play Billing implemented where needed.
- App content declarations, including Data safety, completed.
- Production access application answered with specifics from the test.
Frequently asked questions
Can Gemini CLI publish my app to Google Play?
It can write and run the scripts that upload bundles to a testing track once a service account exists. It cannot open the developer account, fill in Play Console declarations, recruit testers or apply for production access.
How long does a first Google Play release take?
For a new personal account, plan for at least two to three weeks: 14 days of closed testing plus the review of your production access application and the app. Organization accounts can be faster.
Do I need to switch to Antigravity CLI?
That depends on how you access Gemini. Google moved subscriber and free individual accounts to Antigravity CLI in June 2026; access through paid API keys and enterprise plans continues. The release steps in this article are the same with either tool.
Can you handle the Play Console part for us?
Yes. Book a free call and we look at your app and timeline together. Our AI Code Engineers can guide the account setup and take over signing, listing, declarations and review in your own Play Console.
Free, 15 minutes, no obligation
Get your app from agent output to approved release.
Our AI Code Engineers take apps built with Claude Code, Codex, Cursor, Lovable, Base44 and co. through signing, CI/CD, store listings and review – in your own developer accounts.
AI tools, store policies and requirements change quickly – always check the current documentation of the tools and stores you use. This article is independent advice and not affiliated with or endorsed by Apple, Google, Anthropic, OpenAI or any other company or product mentioned; all trademarks belong to their owners.