Field noteNov 18, 20265 min read1 source

Previously limited to generated Cloud Run URLs or shared ai.studio subdomains.

In one minute

Google AI Studio’s publishing documentation describes custom-domain support for apps deployed from Build mode. The feature lets teams put a branded address in front of a managed Cloud Run app, but requires a paid Google Cloud project; the free Starter Tier remains limited to AI Studio subdomains and two apps per project.

What changed: Google AI Studio’s custom-domain deployment guidance adds eligibility and publishing requirements, including Starter Tier limits and paid-project requirements.

What happened

Google AI Studio’s documentation says developers can publish full-stack web apps directly from Build mode. A published app runs as a service on a managed, scalable Cloud Run service and receives a shareable public URL. The documentation describes two project options: an eligible account can use the Google Cloud Starter Tier without setting up its own Google Cloud project or billing account, while a paid deployment requires a Google Cloud project linked to the AI Studio account with Cloud Billing enabled.

The Starter Tier has a limit of two published apps total per project, and those apps must be hosted in a single Cloud Run region. Google says eligibility is not universal. Accounts using Google Workspace, including Education and Nonprofits accounts, may be ineligible, as may accounts with an active or prior paid Cloud billing account, or accounts for which Google does not yet have enough activity evidence. The support team cannot override eligibility; Google directs ineligible users to attach a billing account and publish with a paid project. The publishing flow for eligible users proceeds through the Publish panel, project confirmation, and the free-URL option.

For URLs, AI Studio documents both free ai.studio subdomains and custom domains. The free subdomains are available on both Starter Tier and paid projects, are unique across AI Studio, and are assigned first come, first served. A subdomain is released when its app is unpublished or deleted. Custom domains, by contrast, are available on paid Google Cloud projects. A developer can buy and connect a domain in the Publish panel, or reuse a domain bought through AI Studio. Google says it registers the domain in Cloud Domains, configures Cloud DNS routing, and creates a managed SSL certificate. The app retains its ai.studio subdomain as well. Google says support for transferring existing domains is planned for the future.

The documentation says upgrading a Starter Tier project to a standard Google Cloud project preserves published apps and project resources. The paid tier is positioned for apps needing capabilities beyond Starter Tier, including custom domains, higher quotas, more compute resources, and other Google Cloud products. Domain pricing, setup, subscription management, and quota requests are handled in a separate custom-domain guide. Google’s security guidance assigns responsibility for Cloud Run infrastructure to Google, while the app’s code and dependencies—including AI Studio-generated code and packages—remain the developer’s responsibility. It recommends keeping dependencies updated, republishing after updates, and responding to security notifications.

What the archive adds

New Runtime’s earlier coverage includes Google AI Studio’s move of Gemini API documentation into the workspace only by broad product association, not by the same deployment change. This is a new fact about app publishing, not a restatement of that earlier subject.

Trend signal

This does not yet fit a pattern New Runtime tracks. The documentation describes a useful hosting and domain option, but it does not establish a broader shift in how agents, software interfaces, or AI-native teams are built.

Why it matters

A branded domain is often part of the difference between a prototype that is easy to share and an app that can be shown to customers or colleagues as a recognizable service. Putting domain registration, DNS routing, and certificate provisioning in the AI Studio publishing flow can remove several setup steps from the path between building an app and sharing it. Teams can keep the app on managed Cloud Run while avoiding a generated-only address.

The trade-off is a clearer divide between experimentation and deployment. Starter Tier offers a simple route for eligible accounts, but its app count, region, and domain restrictions make it unsuitable for some public-facing uses. A paid project enables custom domains and higher quotas, but introduces billing and the associated Google Cloud project setup. A custom address does not transfer operational responsibility for app code: teams still need to maintain dependencies and republish fixes. Builders should treat the feature as a simpler publishing path, not as a substitute for production review, security work, or decisions about scaling and service requirements.

Limits and uncertainty

Google’s documentation establishes the supported URL options and project requirements, but it does not give a launch date for custom-domain support. It points readers elsewhere for domain pricing, setup details, subscription management, and quota requests, so the cost and practical quota limits cannot be assessed from this publishing page alone. Google also says Starter Tier eligibility depends on account circumstances and may not be overridden by support. The documentation does not report adoption, reliability, or performance measurements for apps published through this flow. Teams should confirm current pricing, eligibility, regional needs, and service compatibility before choosing a deployment tier.

Skill to build now

Before publishing an AI Studio app, write down the audience, URL requirement, service dependencies, and expected operating needs. If the app is a temporary prototype and the account qualifies, test the Starter Tier’s free ai.studio URL and verify that its single-region and two-app limits are acceptable. If the app needs a branded domain, plan around a paid Google Cloud project, check domain pricing and quotas, and test the URL and certificate before sharing it widely. In either case, add a dependency-update and republishing step to the team’s release checklist; Google explicitly leaves app code and dependencies to the developer.

Reader takeaway

AI Studio can publish a Build-mode app on Cloud Run under a custom domain, but that option belongs to a paid Google Cloud project. Starter Tier is capped at two apps in one region and uses free ai.studio URLs. Choose the tier based on the app’s audience and requirements, and remember that managed hosting does not remove the developer’s responsibility for code and dependencies.

Evidence / sources

  1. 1ai.google.dev/gemini-api/docs/aistudio-deploying
All Field Notes

Next reads

Read next

Open the graph
  1. 01topicGoogleExplore the google topic hub.
  2. 02archiveField NotesOpen the latest editorial analysis.
  3. 03sourcesSourcesSee the original sources this archive cites.
Who read this page?Machine requests, hidden until opened

Loading the privacy-safe route aggregate...

Open the JSON contract