Skip to content

What is Google AI Studio? Free use, getting an API key, limits, building apps and Turkish support

In Google AI Studio you write prompts to Gemini models in the browser and test the results, build apps by describing them, or get a Gemini API key to use in your own app. Using it is free; the free tier covers only some models and limited usage. Turkey is on the official region list, and you must be 18 or older. The key difference: on the free tier, what you send may be used to improve Google products; on billed (paid) usage it is not.

What it does
Lets you try Gemini, build apps and get an API key
Turkey
Turkey is on the official region list; age 18 or older
Turkish
Turkish is available in audio models; official sources don't cover text and the interface
Data
On the free tier, prompts may be used for product improvement
Where
Web (aistudio.google.com)
Free?
Yes; the free tier covers only some models and limited usage
Price
Paid tier is pay-as-you-go (USD)

What does Google AI Studio do?

Google AI Studio is a browser page for developers and enthusiasts who want to work with Gemini models. You type into the chat box to try models, adjust behavior with a "system instruction" (a setting that gives the model a standing role and rules), and get the code for a request you like in one click.

The terms of service describe it as being for "professional or business use by developers", not consumer use. If you want an everyday chat assistant, the same company's consumer app, Gemini, is a better fit.

For everyday chat and an assistant that works with Google products like Gmail and Drive, see the Gemini guide.

For the Google tool that produces interface designs from a description, see the Stitch guide.

PlaygroundTry models in the browser, design system instructions and prompts, and get ready-made code.Web · Personal Google account, age 18 or older, supported region; free · details
App building (Build mode)Build, share and publish a web or Android app by describing it.Web · Building can be tried on the free tier; publishing needs the Starter Tier or a billed Cloud project · details
Gemini API keyConnect your own app, automation or coding tool to Gemini.Web, Terminal · The key is free; limited on the free tier, pay-as-you-go on the paid tier · details

How to get started: account, age and region requirements

  1. Go to aistudio.google.com and sign in with your Google account; no installation is needed.
  2. You must be 18 or older and have verified your age in your Google account. The official page lists these two conditions among the reasons access can be denied.
  3. The country or region you live in must be on the supported list. Turkey is on the list.
  4. Accept the terms of service. New users get a default project and API key created automatically after accepting the terms.

Official sources don't say whether phone verification or ID is required. You can also sign in with a Google Workspace (work or school) account, but the Starter Tier publishing allowance in Build mode is not given to Workspace accounts. If you can't create an API key on an organization account, the official page says the problem is a missing permission; ask your admin for a role or use a personal Google project not tied to the organization.

The official page lists region restrictions, the 18+ age requirement and missing age verification in your Google account as reasons AI Studio won't open. No official page says a VPN is a solution; don't try to get around the region rule.

How do you get a Gemini API key?

An API key is the secret password your own app or an automation tool (such as Make, n8n or a coding assistant) presents to use Gemini. You get the key from AI Studio for free; how much you can use is still limited by your account's tier.

  1. Open aistudio.google.com/apikey. If you are a new user, a default project and key are already created once you accept the terms.
  2. If you have another Google Cloud project: in the left panel go to Dashboard, then Projects, and add the project with "Import projects"; then create a new key in that project from the API Keys page.
  3. Copy the key. Storing it as an environment variable named GEMINI_API_KEY is recommended; the official client libraries read this variable automatically.
  4. Don't write the key into your code, put it in a Git repository, or embed it in a web page or mobile app. A key sitting in code that users can see can be extracted; set up an intermediate layer on your own server instead.
Environment variable outside Windows (Linux/macOS)› export GEMINI_API_KEY=<YOUR_KEY>

Since May 2026, new keys created in AI Studio are created as "authorization keys" (keys tied to a service account); the Gemini API rejects older unrestricted keys. If a key leaks, first create a new key and switch your app to it, then disable the old one.

Don't share your key with anyone or paste it into chat windows. When you give a key to a coding assistant or automation tool, check separately how that tool handles your data.

How do you use Google AI Studio?

  1. Open AI Studio; a new chat prompt is ready.
  2. Open the "Run settings" panel at the top right; write the model's role and rules in the "System Instructions" box.
  3. Type your question in the message box at the bottom and run it with "Run". If you don't like the answer, change the instruction and try again; the instruction can be changed even after the chat has started.
  4. When you are happy with the prompt, get the code in your chosen language with "Get code", or save and share the prompt.

In a chat, every message is resent as part of the prompt. As the conversation grows, the prompt grows and you can hit the model's token limit; for a long task, starting a new chat is cheaper and faster.

In the Run settings panel you also change model settings, safety settings and tools: structured output, function calling, code execution and grounding with search (the model backing its answer with Google Search) are turned on here. On the free tier not all of these tools are guaranteed to be available; the pricing page marks most tools as "not available on the free tier".

System instruction: customer support assistant"You are the customer support assistant for a coffee shop called Bayside Coffee. Only talk about products, orders and shipping. Keep answers to at most 3 sentences in a friendly tone. Don't make up anything you don't know; say "our team will get back to you.""
Prompt: social media copy"Write 3 different Instagram post captions from the product description below. Each should be 2 sentences: one witty, one informative, one a promotion announcement. Product: Handmade cold-process olive oil soap, 100 g."
Prompt: document summary"Summarize the attached contract point by point. Put the term, payment terms and termination conditions under separate headings. Don't make up anything you don't understand; write "not clear in the text.""

Don't upload documents containing customer data, contracts or personal information on the free tier. On free usage, content may be used for product improvement and human reviewers may read it. Details are in the "Data and privacy" section.

Build mode: building an app from a description

Build mode (aistudio.google.com/apps) builds you a working app when you describe what you want in plain language. A new web app gets a React frontend and a Node.js server side; if you like, you can also choose a native Android app (Kotlin and Jetpack Compose) from the platform selector. A live preview is on the right, and the generated code is in the "Code" tab.

Describe itType what you want in the box; if you have no idea, the "I'm Feeling Lucky" button generates a project idea.
Preview and changeFix things through chat, or by marking a spot in the interface (annotation mode) and saying what you want; you can also edit the code by hand.
Share or publishShare it, sync with GitHub, download a ZIP or publish to Cloud Run.
Gemini calls made by people using an app you share count against your usage limit; if you use a paid model, the cost is billed to you.
Is the API key safe?

AI Studio keeps the app's Gemini key as a secret on the server side; it is not in the browser code and others can't see it in a shared app. You can see the key in the Secrets section under Settings. Apps built before May 2026 are moved to this method when you change their Gemini features.

Publishing and allowance

With the Google Cloud Starter Tier, eligible accounts can publish up to 2 apps without opening a billing account. Workspace accounts, accounts that previously had paid Cloud billing and accounts where Google doesn't see enough account history may not be eligible; the support team can't change this. Otherwise a billed Cloud project is required.

Database and sign-in

If you want, the agent can set up Firebase Firestore (data storage) and Sign in with Google; AI Studio also sets up connections to Workspace services such as Gmail, Sheets and Drive. If you collect data from real users, the legal responsibility is yours as the app owner.

Team work

There is no real-time co-editing. To work with your team, connect the app to a GitHub repository; there is two-way sync and a conflict resolution window.

The official page doesn't say which plan unlocks Android app building. If Android doesn't appear in the platform selector, the source doesn't give the reason; check Google's help pages or Build mode's own warnings. An app can be removed for a policy violation (such as malware, impersonation or hate speech), and repeated violations can lead to loss of access.

Which model? Gemini models and their differences

AI Studio has many models, and each can be tried separately. You choose the model from the model selector next to the chat. The "Try it in Google AI Studio" links on the pricing page also open with the model name selected (model=… in the URL). The table below is taken from the descriptions on the pricing and model pages.

ModelWhat it is for, per GoogleOn the free tier?Paid price (per 1 million tokens, USD)
Gemini 4 ArgonFlagship reasoning model: complex software, advanced math and science, multi-step agents, long generation up to 1 million tokensNo; not included in the Google AI Pro plan eitherInput $2.00 · output $10.00 (until November 30, 2026); from December 1, 2026 input $4.00 · output $20.00
Gemini 3.1 Pro (preview)Multimodal understanding, agentic capabilities, "vibe coding"NoInput $2.00 · output $12.00 (prompts up to 200K tokens); longer prompts $4.00 · $18.00
Gemini 3.8 FlashThe smartest Flash model: long-running software work, autonomous agents, complex workflowsYesInput $0.75 · output $3.75 (until December 31, 2026); from January 1, 2027 $1.50 · $7.50
Gemini 3.6 FlashPrevious-generation Flash: speed and multimodal, general and everyday tasksYesInput $0.75 · output $3.75 (until December 31, 2026); from January 1, 2027 $1.50 · $7.50
Gemini 3.5 Flash-LiteThe fastest and cheapest 3.5 model: high-volume tasks, translation, simple data processingYesInput $0.30 · output $2.50
Gemini 3.1 Flash-LiteClose to larger models in quality, much cheaper: high volume, translation, simple data processingYesInput $0.25 (audio $0.50) · output $1.50
Nano Banana 2.1, 2, 2 Lite, ProImage generation and editingNoSeparate tables on the pricing page (by image resolution)
Veo 3.1Cinematic video generation, with audioNoPer second: standard $0.40 (720p/1080p), fast version $0.10 (720p)
Gemini 3.8 Flash TTS and Flash-Lite TTSText to speech (130 and 101 languages)YesFlash TTS: text input $0.50 · audio output $9.00 (until December 31, 2026)
Gemini 3.8 Live, 3.5 TranscribeLive voice conversation and transcriptionYesSeparate line on the pricing page, by audio minute and token
Which one to choose?
  • Most testing and app building: Flash models (available on the free tier).
  • Lots of small tasks done cheaply: Flash-Lite models.
  • The hardest reasoning, complex software and agent work: Argon (paid only).
  • Voice work: the TTS, Live and Transcribe models.
Effect on your usage allowance

Each model has its own limit; on the free tier only models marked "free" work. Preview models have tighter limits and can be shut down.

Version warnings

The models page says only accounts that used the 2.5 family before can access it, and recommends 3.5 Flash-Lite or 3.8 Flash for new projects; the 2.0 family has been shut down. The models page doesn't list Argon, but the pricing page and model card do; this guide is based on those two.

The model list and prices were taken from the official model, pricing and audio pages on October 14, 2026; they change often. The official page doesn't give model speeds as numbers, so there is no "speed" column.

Is Google AI Studio free? Free and paid tiers

Yes. Using AI Studio is free; the free tier gives "limited access to certain models" and free input and output tokens. Your request limits are shown in AI Studio based on your account and the model you choose. Official pages list prices in US dollars.

Free$0
  • Limited access to certain models
  • Free input and output tokens
  • AI Studio access
  • Prompts not used for product improvement
  • Context caching
  • Batch API
  • The most advanced models
Paid (billed)Pay as you go USD
  • Higher rate limits
  • Context caching
  • Batch API (50% cost reduction)
  • Access to the most advanced models
  • Prompts are not used for product improvement
Google AI Pro / UltraSubscription; price not in this source
  • Higher quota than free in the AI Studio interface
  • Does not cover API key usage
  • Does not cover the Deep Research and Antigravity agents
EnterpriseContact sales
  • Dedicated support channel
  • Advanced security and compliance
  • Provisioned throughput
  • Volume discounts
  1. To move to the paid tier, click "Set up billing" in AI Studio and choose or create a Cloud billing account.
  2. New accounts on the prepaid plan may be asked to load at least $5 of credit (or the equivalent in another currency); some accounts are offered a choice of prepay or postpay.
  3. Once billing is linked, Tier 1 starts; as spending and account age grow, you move to Tiers 2 and 3 automatically.

AI Studio usage stays free as long as you don't connect a paid API key. If you connect a key from a paid project, your AI Studio usage with that key is billed too; you can switch between free and paid projects by switching keys.

See it as a table: tier comparison
TierRequirementWho it is for
FreeActive project or free trialTesting, small projects
Tier 1Link an active billing accountApps going into production
Tier 2$100 paid + 3 days since first successful paymentHigher volume
Tier 3$1,000 paid + 30 days since first successful paymentHighest volume

Price and tier information was taken from the official pricing and billing pages on October 14, 2026. The official source doesn't say whether a tax such as VAT is added; check the amount on the payment screen.

Official pricing page last checked on .

Usage limits and tiers: how much per request?

Limits are counted on three measures: requests per minute (RPM), input tokens per minute (TPM) and requests per day (RPD). Exceeding any one of them returns an error. Limits apply to the project, not the API key; the daily request allowance resets at midnight Pacific time. Limits vary by model and are tighter for preview and experimental models.

The official documentation doesn't give exact per-model numbers: it says "limits are shown on the Rate limit page in AI Studio based on your account" and that "specified limits are not guaranteed". These numbers are only visible after you sign in; check them on your own account.

TierHow to qualifySpend cap per 10 minutesBilling tier ceiling
FreeActive project or free trialNoneNone
Tier 1Link an active billing account$10$250
Tier 2Pay $100 + 3 days since first successful payment$50$2,000
Tier 3Pay $1,000 + 30 days since first successful payment$200$20,000 - $100,000+

Moving from free to Tier 1 is usually instant; later upgrades take effect within about 10 minutes. Tier 2 and 3 requirements are based on total Google Cloud spending on the billing account linked to your project. There is an official request form for more quota; an increase is not guaranteed.

If you hit the spend cap you get a 429 RESOURCE_EXHAUSTED error. Fix: wait a bit, retry, reduce expensive requests. On a prepaid account, if credit drops to 0, all keys of all projects linked to the account stop together with a 402 error; auto-reload is recommended.

Using it from Turkey and in Turkish: what is and isn't there?

Turkish replies, a Turkish interface and Turkish audio are separate topics; official sources cover each separately.

Access from TurkeyYesTurkey is on the list of regions where AI Studio and the Gemini API are available. Age 18+ and age verification in your Google account are required.
Turkish prompts and replies in textUnclearThe model pages don't give a language list for text models. To test, write in Turkish in the Playground.
Turkish text to speechYesTurkish is in the official language table of the text-to-speech models (marked for both models).
Live conversation (Live)YesTurkish (tr) appears in the Live API's language list.
Speech transcriptionYesTurkish (tr-TR) appears in the transcription model's language list.
Turkish interfaceUnclearThe official documentation has no line about the AI Studio interface language.
Local-currency price and taxNoThe official pricing and billing pages show only USD; there is no tax line. Check the amount on the payment screen.

Free and paid tier rules work the same for Turkey; for the European Economic Area, Switzerland and the United Kingdom, paid rules also apply to free usage. Turkey is not one of these three regions.

Is what I type used for training? Free vs. paid

This is the most important difference: Google handles your prompts differently depending on your tier. The summary below is from the Gemini API Additional Terms of Service; it is not legal advice.

Unpaid services
  • Direct use of AI Studio and the free API quota count as "unpaid services".
  • The content you submit and the generated responses are used to improve Google products and machine learning technologies.
  • Human reviewers may read inputs and outputs for quality; the data is said to be shown disconnected from your Google account, API key and project.
  • Don't submit sensitive, confidential or personal information.
Paid (billed) services
  • It is stated in writing that your prompts and responses are NOT used for product improvement; they are processed under the data processing addendum.
  • To prevent abuse, prompts and responses are logged for a limited time (the usage policies page gives this as 55 days).
  • Your AI Studio access counts as a "paid service" if you have access to a billed Cloud project or are on a Workspace enterprise account.
  • AI Studio states that when the account has at least one billed API project, it also handles chat prompts under the paid service rules.
EU, Switzerland, United Kingdom

In these three regions, paid-service data rules apply to everything, including AI Studio and the free quota. Turkey is not on this list.

Logs and datasets

In billed projects you can turn on API logs if you want (kept for up to 55 days by default; 7/14/28/55 days can be chosen). If you share logs with Google, they fall under the unpaid service rules.

Don't try work containing customer information, health data, contracts or company secrets on the free tier. Move to a paid, billed project first, or strip the data out.

Commercial use and rules: what can and can't you do?

Source: the official Gemini API Additional Terms of Service. This is not legal advice; for a commercial project, read the full terms.

Allowed
  • AI Studio and the API are defined for professional or business development
  • Google does not claim ownership of generated content; responsibility for its use is yours
  • You can offer the app you build to users in supported regions
Not allowed
  • Apps aimed at, or likely to be accessed by, people under 18
  • Developing models that compete with the service, reverse engineering
  • Clinical use or medical advice
  • Circumventing safety measures, violating the prohibited use policy
  • Using the free quota in an app offered to users in the EU, Switzerland or the United Kingdom

Google states that it may generate the same or similar content for others; in other words, a model output may not be unique to you. Review generated code yourself before using it.

Common problems and fixes

SymptomFix
AI Studio won't open or shows "not available in your region"Check the region list, the 18+ age requirement and age verification in your Google account. A VPN is not recommended as a fix.
429 RESOURCE_EXHAUSTED errorWait a bit, retry, reduce expensive requests; if it keeps happening, use the limit increase form.
402 Payment Required error (prepaid account)Credit dropped to 0 and all keys linked to the account stopped. Load credit from the AI Studio billing page; turn on auto-reload.
An app I shared returns 403 Access RestrictedTurn off privacy extensions, or in Build mode tell the agent "fix the build issues in the current code" and share again.
The "Create API key" button is disabled ("no permission to create keys in this project")You don't have permission to create keys in the project; ask your admin for a role such as Project Editor, or open a new project not tied to the organization.
My old key is rejectedThe Gemini API rejects unrestricted standard keys. Create a new key in AI Studio (since May 2026 new keys are created as authorization keys).
The publish button doesn't offer the Starter TierWorkspace accounts, accounts that had paid Cloud billing in the past and accounts with insufficient history are not eligible; support can't change this. Link a billing account and use standard publishing.

Leaving the paid tier, refunds and removal

  1. Going back to the free tier Disable Cloud billing on the projects you want to move down to the free tier; the project drops to the free tier.
  2. Prepaid credit Credit you load expires after 12 months and is not refunded. If you switch to a postpay account, the remaining credit is refunded to your payment method. If you close the billing account for another reason, the remaining credit is lost.
  3. Postpay account Refunds are subject to Google Cloud's refund policy.
  4. An app you published If you remove a Starter Tier deployment from Build mode, your ai.studio subdomain is released for others to take.
  5. Google AI Pro / Ultra This subscription is managed separately; the cancellation steps are not in this source. Check the Google AI plans page.

Google AI Studio vs. ChatGPT vs. Claude

AI Studio is not a chat app; it is a developer workspace. If you want a ready-made assistant, use Gemini, ChatGPT or Claude; if you want to build apps with Gemini models and get an API key, use AI Studio.

Frequently asked questions

What is Google AI Studio?

Google's developer workspace where you try Gemini models in the browser, design prompts, build apps by describing them and get a Gemini API key.

How do I get a Google AI Studio API key?

From aistudio.google.com/apikey. New users get a default project and key created once they accept the terms; the key is free, and usage is subject to free-tier limits.

Is Google AI Studio free?

Yes. Using AI Studio is free; the free tier has limited access to certain models and free tokens. If you connect a paid API key, usage with that key is billed.

What are Google AI Studio's limits and when do they reset?

Limits are counted as requests per minute, tokens per minute and requests per day, and they apply to the project. The daily request allowance resets at midnight Pacific time. Per-model numbers appear on the Rate limit page in AI Studio based on your account.

Does Google AI Studio support Turkish?

Turkish is officially listed for the audio models (text to speech, live conversation, transcription). Official sources don't give a language list for the text models or the interface; test it by writing in Turkish in the Playground.

Can Google AI Studio be used from Turkey?

Yes; Turkey is on the official list of available regions. You must be 18 or older and have verified your age in your Google account.

Can you build apps with Google AI Studio?

Yes; in Build mode you build a web app (React and Node.js) or an Android app by describing it. Publishing needs the Starter Tier (up to 2 apps) or a billed Cloud project.

Can Google AI Studio do text to speech?

Yes; there are text-to-speech models and Turkish is marked in their language table. Per the pricing page some are on the free tier, and paid usage is priced per audio output token.

Can Google AI Studio generate video?

The model list includes video generation models; per the pricing page, video models are not on the free tier and are priced per second on the paid tier.

Is what I type used for training?

On free usage, yes: content may be used to improve Google products and human reviewers may read it. On billed (paid) usage it is not. This is not legal advice.

Google AI Studio or ChatGPT?

AI Studio is not a ready-made chat app; it is a developer workspace. For a ready-made assistant use Gemini or ChatGPT; to build apps and get an API key use AI Studio.

How do I cancel a Google AI Studio subscription?

AI Studio has no subscription of its own. To leave the paid tier you disable Cloud billing on the project; if you have a Google AI Pro or Ultra subscription, that is managed separately.

Related guides

Sources

Not clear in the official sources: Per-model request and token limits: the official documentation gives no numbers; they appear on the Rate limit page of your account. This guide only has the tier and spend cap table.,Turkish support in the text models and the interface language: official sources don't say. To test, write in Turkish in the Playground.,The plan requirement for Android app building and any phone verification/ID requirement: official sources don't say. Check the on-screen warnings in Build mode and when creating an account.,Local-currency price and tax: prices are listed only in USD. Check the amount on the payment screen.,Difference between the models page and the pricing page for the newest model: the pricing page and model card were used.