Google AI Studio

Experiment with Gemini prompts

Google AI Studio is Google's browser workspace for trying Gemini models and developing applications around them. Its Playground supports chat prompts with multiple turns, system instructions and examples. The quickstart shows a prompt being refined through repeated questions and changes to the instructions.

For everyday brainstorming or a first draft, Gemini Apps provides the conversational assistant. AI Studio instead gives developers a place to experiment with prompts and move them into API-backed applications.

Run settings provide model parameters and safety settings, alongside tools such as structured output, function calling, code execution and grounding. These options serve different experiments; their presence does not establish that every model supports every configuration. Once a prompt is ready, the documented Get code action helps move it into a chosen programming language using the Gemini API.

Build an application and inspect its code

Build mode takes a natural-language description and generates an application with a live preview. Google's current guide describes web applications with a React frontend by default and a Node.js server runtime. It also describes native Android projects using Kotlin and Jetpack Compose, with a browser-based emulator for previewing them.

The generated code remains available for inspection and editing. A developer can refine the project through chat, import from GitHub, connect a repository for two-way sync or download a ZIP for local development. The guide describes app sharing and Cloud Run deployment; publishing a working URL still leaves the developer responsible for the app's behavior and data.

Keep API keys, sharing and collaboration distinct

For generated Gemini integrations, the Build guide describes storing the API key as a server-side secret. Shared or deployed apps use the owner's key for their Gemini calls. Sharing therefore affects the owner's usage limits and can incur costs with paid models, even when another person is using the app.

The documentation has qualifications worth checking before relying on team features. It describes edit-permission sharing but also says direct real-time collaborative editing is unsupported. Repository collaboration is a documented route. It likewise describes automatic Firebase provisioning while another FAQ says direct storage support is being added. Confirm the intended setup instead of treating either statement as a universal storage guarantee.

Check the project and billing tier

The current billing guide says AI Studio remains free unless a paid API key is linked for paid features. Free-tier access covers certain models within their rate limits. Paid usage, deployment services and app sharing should be assessed separately.

AI Studio exposes usage and billing controls. The billing guide warns that processing delays can allow spending beyond a configured cap. Start with an appropriate project, check the selected key and plan, and review the generated application's secrets and access before sharing it. This entry describes official documentation; it does not claim a logged-in build or deployment test.