Gemini prototyping
Test chat, realtime, video, structured-output, tool-use, and multimodal workflows before writing a full integration.
Independent tool overview
Google AI Studio at a glance
Google AI Studio is Google's browser-based workspace for testing Gemini models, configuring prompts and tools, obtaining API keys, and generating deployable web or Android applications.
Visit the official Google AI Studio site ↗
- Developer
- Core platform cost
- Free
- Models
- Gemini and Gemma
- App targets
- Web and Android
- API
- Gemini Developer API
- Deployment
- ai.studio or Cloud Run
Overview
What Google AI Studio is
Google AI Studio has grown well beyond a prompt playground. Developers can compare Gemini behavior, tune run settings, test multimodal input, enable tools, inspect token use, get integration code, and move a prototype into the Gemini Developer API.
Build mode can now generate full-stack web applications with a React frontend and Node.js server runtime, or native Android apps with Kotlin and Jetpack Compose. It supports live previews, multi-file editing, server-side secrets, npm packages, Google Workspace integrations, GitHub sync, ZIP export, and deployment to Cloud Run or an ai.studio subdomain.
AI Studio itself is free in supported regions, but the cost and data treatment change when a project uses a paid Gemini API key, paid model, external service, or Cloud Run deployment. Shared applications consume the owner's API quota, so public usage needs authentication, budgets, rate limits, and abuse controls.
Use cases
Who Google AI Studio is best for
The strongest fit depends on the job you need the product to complete, not the size of its feature list.
API onboarding
Create a Gemini API key, validate a prompt, inspect settings, and export starter code in a supported language.
Full-stack AI apps
Generate and iterate on React and Node.js applications with server-side secrets, packages, and external APIs.
Native Android prototypes
Build Kotlin and Jetpack Compose projects, preview them in a browser emulator, and test on a device.
Model and cost experiments
Compare model quality, context, latency, token usage, tool behavior, and free-versus-paid limits on a representative workload.
Capabilities
Core Google AI Studio features
Prompt and model playground
Runs multi-turn prompts against supported Gemini models with adjustable model, safety, sampling, and output settings.
Multimodal testing
Supports workflows involving text, images, audio, video, files, realtime streaming, and model-specific generation capabilities.
Developer tools
Enables structured output, function calling, code execution, grounding, token inspection, API-key creation, and Get code exports.
Full-stack Build mode
Generates a React client and Node.js server, manages multiple files, installs npm packages, and shows a live app preview.
Native Android generation
Creates Kotlin and Jetpack Compose applications with a browser emulator, device installation, and internal Play testing paths.
Server-side secrets
Stores Gemini and third-party credentials in the server environment instead of embedding them in browser code.
Workspace integrations
Can connect generated apps to Gmail, Sheets, Docs, Drive, Calendar, and other Google Workspace APIs with handled OAuth setup.
Source and deployment options
Supports GitHub two-way sync, ZIP download, sharing, ai.studio publishing, and Cloud Run deployment.
Process
How the Google AI Studio workflow works
Step 1
Define the test case
Choose a real user task, expected schema, sensitive-data boundary, quality rubric, latency goal, and maximum cost.
Step 2
Prototype the prompt
Select a model, add representative inputs, configure tools and safety, and inspect several runs for variance.
Step 3
Measure usage
Count input and output tokens, test rate limits, and compare free and paid model economics before scaling.
Step 4
Build or export
Use Get code for an API integration or Build mode for a web or Android application.
Step 5
Secure the app
Keep credentials server-side, validate user input, restrict tools, protect connected accounts, and add quotas and abuse prevention.
Step 6
Deploy with controls
Use GitHub and your normal review process, configure budgets and monitoring, test failures, and only then publish through ai.studio, Cloud Run, or another host.
Cost
Google AI Studio pricing and free plan
Google says AI Studio usage is free in supported regions. The Gemini API has a limited free tier and model-specific paid token pricing. As one current example, Gemini 3.7 Flash is promoted through December 31, 2026 at $0.75 per million input tokens and $3.75 per million output tokens. Cloud Run and third-party services are billed separately.
Google AI Studio
Free
Use the browser workspace in supported regions without a separate AI Studio subscription.
- Prompt testing
- Build mode
- API-key setup
- Usage remains subject to quotas and availability
Gemini API Free
$0 within limits
Limited access to eligible models with free input and output tokens.
- Lower rate limits
- Not every model or tool is included
- Free-tier content may be used to improve Google products
- Project-level quotas apply
Gemini API Paid
Model-specific usage
Higher limits, advanced models, caching, batch discounts, and paid tools.
- Gemini 3.7 Flash promo: $0.75/M input
- Gemini 3.7 Flash promo: $3.75/M output
- Paid-tier content not used to improve products
- Prices differ by model, modality, tool, and service level
Enterprise and hosting
Custom + infrastructure usage
Enterprise support and Cloud Run deployments have separate commercial terms.
- Provisioned throughput options
- Volume pricing may be available
- Cloud Run usage billed separately
- External APIs and storage may add cost
Pricing checked . Check current pricing at the source ↗
Assessment
Google AI Studio strengths and limitations
Where it stands out
- Fast path from Gemini experiment to working code
- Free browser workspace with a Gemini API free tier
- Broad multimodal and tool-testing capabilities
- Full-stack web and native Android generation
- Server-side secrets reduce accidental browser key exposure
- GitHub, ZIP, Workspace, and Cloud Run integration paths
- Direct access to Google's latest developer models
What to consider
- Generated applications still require code review, testing, security work, and operational ownership
- Shared apps consume the creator's API quota and can create unexpected costs
- Free-tier prompts and outputs may be used to improve Google products
- Model, preview-feature, quota, and pricing availability changes frequently
- Cloud Run, storage, databases, Google APIs, and third-party services can add separate charges
- Generated code may contain insecure assumptions, broken dependencies, or weak error handling
- Users with access to a shared app may be able to view its code and fork it
- AI Studio is optimized for the Gemini ecosystem rather than broad multi-provider evaluation
Compare
Google AI Studio alternatives
The right alternative depends on the specific output, workflow, controls and budget your project requires.
Coding
Replit Agent
Choose Replit Agent for a broader hosted coding, database, deployment, and collaboration environment beyond the Gemini ecosystem.
Explore Replit Agent →Design
Vercel v0
Choose Vercel v0 when polished web interfaces and tight Vercel deployment integration are the priority.
Explore Vercel v0 →Coding
Genkit
Choose Genkit for a code-first, open-source framework to build and evaluate production AI flows with Firebase and Google Cloud integrations.
Explore Genkit →Questions
Google AI Studio FAQs
What is Google AI Studio?
It is Google's browser workspace for testing Gemini models, configuring prompts and tools, creating API keys, exporting code, and building web or Android apps.
Is Google AI Studio free?
Google says AI Studio usage is free in supported regions. Paid Gemini API calls, Cloud Run, and connected services can still create charges.
Can Google AI Studio build full-stack apps?
Yes. Build mode can generate a React frontend and Node.js server runtime with secrets, npm packages, external APIs, and live preview.
Can it build Android apps?
Yes. AI Studio supports Kotlin and Jetpack Compose projects with a browser emulator, device installation, and an internal Play Store testing path.
Is my Gemini API key exposed when I share an app?
For current Build-mode apps, Google says the Gemini API key is stored as a server-side secret and not included in client-side code. Users still need to protect other secrets and review exported code.
Who pays when other people use my shared app?
The app's Gemini calls count against the API key and project configured by the app owner, so the owner needs budgets, quotas, authentication, and monitoring.
Does Google use AI Studio data to improve its products?
Google's pricing table marks Free-tier content as used to improve products and Paid-tier content as not used for that purpose. Review current terms for your project and region.
Bottom line
Our Google AI Studio verdict
Google AI Studio is one of the fastest ways to test Gemini and turn a successful experiment into code or a functioning app. Its new full-stack and Android capabilities are substantial, but generated software is still software: teams must review security, privacy, reliability, billing, and deployment behavior before inviting real users.
Visit Google AI Studio website ↗
AI training for the future of work.
Get access to all our AI courses, hundreds of real-world AI use cases, live expert-led workshops, an exclusive network of AI early adopters, and more.
AI Courses
Get unlimited access to all of our current & upcoming industry-specific AI courses for the duration of your subscription.
Daily Guides
To keep up with the rapid pace of AI, our team publishes AI implementation guides daily. Our library contains 300+ practical use cases to automate real-world work.
Workshops
Join weekly, live, interactive sessions with industry leaders who are at the forefront of AI for hands-on implementation guidance and exclusive insights.
Community
Network with an exclusive community of AI-first professionals who are working smarter with AI. Learn how early adopters are using AI in their work and businesses.