A place to experiment with Gemini
Google AI Studio is the developer-facing service for trying Google’s Gemini models in a browser. You enter prompts conversationally, compare responses across models, and test multimodal capability by supplying images or audio.
From the settings you tried, you can see the code for calling the Gemini API. Without setting up a development environment first, prompt experiments, obtaining an API key, and building it into your own app all start from one screen.
How it differs from the Gemini app
The Gemini app is for everyday questions, writing, research, and image generation through a conversation window. Google AI Studio is where you work out which model to use, how to phrase the instructions, and what output format to require.
Use the Gemini app to consult Gemini; use AI Studio and the Gemini API to build something with Gemini or to run the same prompt repeatedly. It looks developer-only, but it also works as a place to experience the differences between models and what a long context actually does.
What you can do
- Switch between Gemini models and compare responses
- Try prompts containing text, images, audio, or video
- Adjust system instructions and output format
- Obtain an API key for the Gemini API
- Export what you tried as Python or JavaScript code
It is also a common entry point for trying Gemini image models such as Nano Banana through the API. Available models and features change, so check the model names on screen against the official documentation.
Cautions
An API key issued in AI Studio is a secret. Do not commit it to a public repository or embed it in browser JavaScript. When calling from an application, keep it in server-side environment variables or a secrets manager.
What is free, what incurs charges, and how input data is handled vary by model and plan. Check the terms and your data settings before entering personal or confidential material.