Google AI Studio Complete Guide (2026) | Free Gemini API Access & API Key Setup

A lot of people get the Gemini app and Google AI Studio confused, but they serve different purposes. The Gemini app is for using AI; AI Studio is for building and testing AI. The clearest dividing line is the API key: the Gemini app can’t generate one; you create it in AI Studio or the Google Cloud console.

A quick clarification: at the end of July 2026, Google canceled its planned standalone AI Studio mobile app and moved the related features into the Gemini app. The web version of AI Studio (aistudio.google.com) is unaffected and continues to operate, so don’t be misled by claims that “AI Studio is shutting down.”

What Is Google AI Studio

Google AI Studio (aistudio.google.com) is Google’s platform for testing and developing with Gemini models. You can write prompts and see responses directly, set System instructions, tune parameters such as temperature / top-P / top-K, and test structured output (JSON), Function calling, Code execution, and Grounding with Google Search.

Once you’ve finished testing, click “Get code” to export the entire prompt setup as Python, Node.js, or REST and bring it straight into your app. This workflow removes much of the friction caused by the gap between the playground and production.

Core differences between AI Studio and the Gemini app:

The interface is mainly in English. Taiwan is an officially supported region, so you can connect directly without extra setup.

Getting Started

Sign In

Go to aistudio.google.com and sign in with any Google account. You need to be 18 or older. No credit card or waitlist is required.

Run Your First Prompt

After signing in, you land in the Prompt playground. The layout is roughly three sections: a left navigation menu, a center chat area, and a right settings panel.

Ignore the right panel for now. Type any question in the center input box and click “Run” to see the model respond.

Want the model to play a specific role or follow consistent rules? Fill in the “System instructions” field. For example:

You are a technical documentation writer for Traditional Chinese readers in Taiwan. Explain complex concepts with concrete examples.
Always reply in Traditional Chinese and do not use Simplified Chinese.

System instructions are included before every conversation, so you don’t need to repeat them each time.

Tune the Parameters

The most commonly used controls in the right panel are:

• Temperature: Controls how creative the output is. 0 gives the most deterministic, stable output; values above 1 make it more random. Lower it for code or tasks that need consistent formatting; raise it for creative copy or more varied results.
• Top-P: Controls the cumulative probability range of candidate tokens and affects output variety together with Temperature. The default works for most situations.
• Top-K: Limits each step to sampling from the K most likely tokens.
• Model: Switch models from the dropdown. Choose the latest Flash or Pro (Preview, no free API tier); lighter models such as Flash-Lite are also available. Options change quickly; see the official model list.

Key Features

Structured output (JSON): Forces the model to return JSON that matches your specified schema, with no extra explanatory text. Useful for data pipelines because it avoids extra parsing work.

Function calling: Lets the model choose which function to call and what arguments to pass; your app performs the actual execution. This is commonly used for AI agents and tasks that need external tools.

Code execution: The model can run Python code in a sandbox and bring the result back into the conversation. Useful for math and data-processing tasks.

Grounding with Google Search: Lets the model search Google before responding, reducing the risk of outdated answers to questions about events after its knowledge cutoff.

Multimodal input: Upload images, audio, or video and ask questions about them, not just text.

Generative media: Access generative models such as Veo (video generation) and Nano Banana (image generation), depending on your account permissions.

Get code: Export the full prompt setup as Python, Node.js, or REST in one click. This is one of the easiest ways to bridge testing and production.

New after Google I/O 2026: Native Android app generation, one-click Cloud Run deployment (through Firebase), and Workspace connections. Natural-language app generation was already available; I/O 2026 mainly added deployment and integration capabilities, helping non-engineers build deployable apps faster.

Free vs. Paid

A few practical points to keep in mind:

Free tier is fine for testing, not for volume: If you’re just testing prompts or validating features, the free tier is enough. Batch processing or a service with steady traffic can quickly run into 429 (Too Many Requests) errors. If that happens, add retry logic or switch to the paid tier.

Data privacy on the free tier: When you use it for free, prompts and responses may be used by Google to improve its products and may be read by human reviewers (exceptions in the official terms). Don’t put sensitive data — personal information, customer data, credentials, or business secrets — in free-tier prompts. Paid-tier data is not used for training, which matters for commercial use.

See ai.google.dev/pricing for current rates (billing setup: official guide); figures change as models and versions are updated.

Who Should Use AI Studio vs. the Gemini App

A simple way to decide:

The Gemini app is enough if you want to ask questions, chat, write emails, organize information, or work with Google Docs / Gmail. You don’t need to code; just open the Gemini app.

You need AI Studio if you:

• Want a Gemini API key to connect to your own app
• Are testing different System instructions and parameter combinations
• Need Structured output / Function calling / Code execution
• Want to export a prompt as runnable code
• Create technical content and need to demonstrate an API workflow

The key distinction is that AI Studio is a development tool, while the Gemini app is a user interface. If you don’t plan to write code, the Gemini app is enough.

How to Get a Gemini API Key

This is one of the main reasons people come to AI Studio, and the steps are short:

1. Sign in at aistudio.google.com
2. Find “Get API key” in the left menu and click it
3. Click “Create API key.” The system creates a Google Cloud project automatically, or lets you choose an existing one
4. Copy the full key and store it in a password manager or another secure place
5. If you can’t find the key, check the API keys page first. If you can’t retrieve it, create a new one, update your app and confirm it works, then delete the old key (official guide)

Once you have the key, the quickest way to connect (Python) is:

pip install google-genai

from google import genai

client = genai.Client(api_key="YOUR API KEY")
response = client.models.generate_content(
    model="gemini-3.8-flash",
    contents="Explain Function calling in Traditional Chinese"
)
print(response.text)

The example uses gemini-3.8-flash, which is available on the API free tier. Models such as the Pro preview have no free tier, so you’ll need to enable billing first. See the full SDK documentation at ai.google.dev/gemini-api/docs.

Conclusion

AI Studio’s biggest strength is how short it makes the distance from testing to production. Tune a prompt in the playground, confirm the output format, and click “Get code” to get code you can paste into your project. That saves time compared with looking up the API documentation and assembling the calls yourself.

Two things deserve attention. First, the free tier’s quota was cut in December 2025, and 429 errors are noticeably more common. If you’re building a production service, plan to move to the paid tier early; being rate-limited at a critical moment is a problem. Second, free-tier data may be used for training. Treat that seriously if you’re working with commercial or private information.

The English-only interface may be a hurdle for people who prefer a localized UI. The models themselves fully support Traditional Chinese input and output, so you can use Traditional Chinese in your conversations; you just need to get used to the English labels on buttons and settings.

Overall, AI Studio is a low-friction entry point into the Gemini ecosystem for developers and technical content creators in Taiwan. It’s free, available directly in Taiwan, and has no waitlist. Try the playground before deciding whether you need a paid tier.

Further Reading

• Gemini Complete Guide
• Nano Banana Guide
• Gemini Student Plan Guide

---

— Penchan