- Guides
- Tool guides
- How to use Google AI Studio
How to use Google AI Studio
Google's free playground for its newest models, a no-code app builder, and the place you get a Gemini API key. Here is when to use it instead of the Gemini app.
We have no affiliate deal with Google, and nobody paid to appear on this page.
What Google AI Studio is
Google AI Studio is Google's free, browser-based workspace for trying its AI models, building small apps from a prompt, and getting a Gemini API key. It lives at aistudio.google.com. Google lists it as "no cost to start", with paid tiers for higher rate limits and more advanced models 8.
It has three parts you will actually use. The Playground is a chat with every setting exposed: model, system instructions, tools and safety settings 1. Build mode turns a plain-English description into a working web or Android app 2. The API Keys page gives you the key that connects your own code or no-code tools to Gemini 3.
Who it is for: founders and operators who want to prototype an internal tool, test a model before paying for API use, or see exactly how a prompt behaves before handing it to a developer. If you only want an assistant for email and documents, the Gemini app is the better fit.
AI Studio vs the Gemini app
The Gemini app is a finished assistant; AI Studio is the workbench behind it. Both run Google's Gemini models, but they suit different jobs and come with different data terms.
| What you get | Gemini app | Google AI Studio |
|---|---|---|
| Made for | Everyday chat, research, Gmail and Docs help | Testing prompts, building apps, API keys 1 |
| Price to start | Free; Google AI Pro $19.99 a month 8 | Free to start; paid API billing for more 8 |
| Model choice on free use | Auto or Flash-Lite 9 | Gemini 4 Hydrogen, Flash and Flash-Lite free 4 |
| Gemini 4 Argon | Google AI Pro and Ultra 9 | Google AI Pro and Ultra, or a paid API key 6 |
| Controls | Model and thinking level | System instructions, tools, safety settings 1 |
| Builds apps | No | Yes, web and Android 2 |
| Connects to Gmail, Drive | Yes, built in | Only inside apps you build 2 |
| Your data on free use | Used to improve Google's services while History is on; a Chat content setting decides training 11 | Free-tier content is used to improve Google's products 4 |
The privacy point that matters: Google's pricing page says content on the free tier of the Gemini API and AI Studio is used to improve its products, while paid-tier content is not 4. Keep client data and anything confidential off the free tier.
Your first prompt
Open AI Studio, set a system instruction, ask a question, and compare answers as you change the instruction. That loop is what the Playground is for. Google's quickstart walks through it in this order 1:
- Open the PlaygroundGo to
aistudio.google.comand sign in with a Google account. A new chat prompt opens by default. - Pick a modelFor free use, start with Gemini 4 Hydrogen. It is free on the free tier, and Google pitches it for speed and everyday coding and agent work 4 7.
- Write the system instructionClick the Run settings icon at top right and fill in System Instructions. This is the standing brief the model follows on every turn.
- Ask and runType a question a real user would ask in the Type something... box and click Run.
- Tune and compareEdit the system instruction, run the same question again and compare. Long chats count toward the model's token limit, because every message is sent with each prompt.
- Keep itSave the prompt to reuse or share, or click Get code to hand a developer the exact call.
A system instruction you can copy for a first test, written for a sales team:
You are a sales assistant for a B2B software company that sells scheduling tools to clinics. Answer in under 120 words. Ask one clarifying question if the request is vague. Never invent prices, customer names or statistics; say "I don't know" instead. End every answer with one suggested next step.
Then test it with a question such as:
A clinic manager says our tool is too expensive compared with a free calendar. Draft my reply.
The Run settings panel also lets you switch on tools: structured output, function calling, code execution and grounding 1. Grounding is the one most non-developers want, because it lets answers draw on current search results instead of the model's training data alone.
Tip: use AI Studio as a prompt editor. Google's own Gemini team described, in a release episode we covered in October 2026, using the model as a copy editor rather than a ghostwriter. Paste a rough draft and ask: "Sharpen this. Keep my ideas and my voice. Then list the 3 weakest parts, 3 ideas I missed, and 3 things that might confuse a reader." Fix the weak parts yourself.
Building an app without code
Build mode turns a description into a working app you can preview, edit, share and deploy, without writing code. Open the Build tab (aistudio.google.com/apps), describe the app, and a live preview appears beside the generated code 2.
Under the hood, Google says the default web stack is a React front end with a Node.js server. The agent can also set up Firestore for storage and Firebase Authentication, including "Sign in with Google", and connect to Gmail, Sheets, Docs, Drive and Calendar with OAuth handled for you 2. It can also build native Android apps in Kotlin.
- Describe the end stateSay what the finished app does and what "done" means, not the steps. Use the AI chips in the prompt box to add features such as image generation or Google Maps data 2.
- Or start from someone else'sOpen the App Gallery, pick a project and choose Copy App to remix it.
- Refine in chat or by pointingAsk for changes in the chat panel, or switch on annotation mode, highlight a part of the screen and describe the change.
- Ship itDeploy to Cloud Run for a public URL (Google Cloud pricing may apply), sync to GitHub, or download a ZIP 2.
A prompt in the style the newsletter has taught for handing over a whole job at once:
Build a single-page dashboard for our sales data. I will upload a CSV with columns date, region, product and revenue. Add filters by region and month, a line chart of monthly revenue, and a table that flags the 5 biggest month-over-month changes with a one-line explanation for each. Done means it opens in the browser with no errors. If you hit an error, fix it yourself before you stop.
Two things to know before you share a build. Apps are private by default, and anyone you share with can view and fork the code. Shared apps also spend your API usage, and AI Studio warns you first if that could cost money 2. Google keeps your API key server-side as a secret rather than in the page code.
Tip: Build mode is right for internal tools, demos and prototypes. For a product with real users, sync to GitHub and have a developer review it before launch.
Getting an API key
You need an API key only to call Gemini from your own code or from tools such as Zapier, Make or n8n. New users get a default project and key created automatically once they accept the terms 3.
- Open the API Keys pageIn AI Studio, click Get API key.
- Create the keyClick Create API key in the project you want. If an existing Google Cloud project is missing, go to Dashboard, Projects, Import projects first 3.
- Lock it downUse Add restrictions, then Restrict to Gemini API only, so a leaked key cannot be used for anything else.
- Store it safelyNever paste a key into a shared doc, a public repo or client-side code. Set a billing alert in the Google Cloud console.
Google's quickstart sets the key as an environment variable, then makes a first call to Gemini 4 Hydrogen 10:
export GEMINI_API_KEY="YOUR_API_KEY"
curl -X POST "https://generativelanguage.googleapis.com/v1beta/interactions" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-H 'Content-Type: application/json' \
-H "Api-Revision: 2026-05-20" \
-d '{
"model": "gemini-4-hydrogen",
"input": "Explain how AI works in a few words"
}'
To move to the paid tier, click Set up billing on the API Keys or Projects page, link a billing account and prepay at least $5 in credits. Usage then shows under Dashboard, Usage 14.
Limits and pricing
The AI Studio interface is free to use with rate limits; you pay only when you add API billing or a Google AI subscription. Free use covers Gemini 4 Hydrogen and the Flash and Flash-Lite models, but not Gemini 4 Argon, Gemini 4 Carbon or Gemini 3.1 Pro 4.
Rate limits are counted per project, not per key, as requests per minute, tokens per minute and requests per day. Daily quotas reset at midnight Pacific time. Linking billing moves you to Tier 1; spend thresholds move you up to Tier 2 and Tier 3 5. Your current numbers appear on the rate limit page inside AI Studio.
| Model | Free tier | Paid input / output |
|---|---|---|
| Gemini 4 Argon | Not available | $2 / $10 until 30 Nov 2026, then $4 / $20 4 |
| Gemini 4 Carbon | Not available | $2 / $10 until 30 Nov 2026, then $4 / $20; paid API keys only 13 |
| Gemini 4 Hydrogen | Free | $1.50 / $7.50 4 |
| Gemini 3.1 Pro (preview) | Not available | $2 / $12 up to 200K-token prompts 4 |
| Gemini 3.8 Flash | Free | $0.75 / $3.75 until 31 Dec 2026, then $1.50 / $7.50 4 |
| Gemini 3.5 Flash-Lite | Free | $0.30 / $2.50 4 |
Subscribers get more inside AI Studio. Google AI Pro ($19.99 a month) and Ultra (from $99.99) add premium models in the Playground, including Gemini 4 Argon, with higher daily allowances shared across the Playground and Build mode 6 8. The subscription covers the AI Studio web interface only: calls made with an API key are billed separately, and Deep Research and the Antigravity preview need a paid key 6.
Gemini 4 Argon reached the Gemini app for Pro and Ultra subscribers on 13 October 2026 12. Google describes it as its efficient Gemini 4 frontier reasoning model, with a 1 million token context window 7. Gemini 4 Carbon, which Google calls its strongest reasoning model 7, opened on 10 November 2026 to all paying developers through the API in AI Studio, and in the Gemini app to Ultra subscribers 13.
Questions people ask
Is Google AI Studio free?
Yes, to start. You can use the Playground and Build mode free with rate limits, on Gemini 4 Hydrogen and the Flash and Flash-Lite models. Gemini 4 Argon needs a Google AI Pro or Ultra plan in AI Studio, or a paid API key.
What is the difference between Google AI Studio and Gemini?
Gemini is the consumer assistant app, connected to Gmail, Docs and Drive. AI Studio is the developer workspace for the same models, with full control over system instructions and tools, an app builder and API keys.
Does Google AI Studio train on my data?
On the free tier, Google says content is used to improve its products. On the paid tier it is not. Use a paid key for anything confidential.
Can I build an app in Google AI Studio without coding?
Yes. Build mode creates a React and Node.js web app or a native Android app from a description, can add a database and Google sign-in, and deploys to Cloud Run or GitHub.
How do I get a Gemini API key?
In AI Studio, click Get API key, then Create API key in a project. Restrict the key to the Gemini API and store it as an environment variable, never in shared files.
Related
New features, explained the week they ship
195,000 readers get Uncover AI's practical AI briefing. We show you what is new and how to use it in your work.
Sources
- Google AI Studio quickstart
- Build mode in Google AI Studio
- Using Gemini API keys
- Gemini Developer API pricing (updated 22 Oct 2026)
- Gemini API rate limits
- Google AI Pro and Ultra in AI Studio
- Gemini API models
- Google AI plans and pricing (US)
- About Gemini models in the Gemini app
- Gemini API quickstart
- Gemini Apps Privacy Hub
- Gemini app release notes
- Gemini 4 Carbon is now available (Google, 10 Nov 2026)
- Gemini API billing