Google AI Studio
Google AI Studio (aistudio.google.com) is a free, browser-based developer environment for prototyping with the Gemini API. It is Google's self-described “fastest path from prompt to production” — you can design a system prompt, test it with multimodal inputs, and export working Python or JavaScript code in under five minutes, without installing anything.
What It Is
AI Studio is an interactive web IDE built for developers and researchers, not end users. Its purpose is to eliminate the setup friction between “I have an idea for a Gemini-powered feature” and “I have working code I can integrate.” It surfaces the full Gemini API in a visual, explorable interface.
Crucially, AI Studio uses the Gemini Developer API — not the enterprise-grade Vertex AI API. The distinction matters for developers: AI Studio is for prototyping and personal projects. For production deployment with SLAs, HIPAA compliance, VPC isolation, and enterprise IAM, you move to Vertex AI.
Key Features
Prompt Editor
Design system prompts, add few-shot examples, switch between single-turn and multi-turn chat mode, configure structured output schemas, and adjust model parameters (temperature, top-p, top-k, max output tokens) — all with live preview.
Model Selection
Access to the full Gemini model roster — Gemini 3 Pro, Gemini 3 Flash, Gemini 2.5 Pro, Gemini 2.5 Flash, and experimental models. Switch models mid-session to compare responses and costs.
Multimodal Input
Attach text, images, video clips, audio files, and documents directly in the prompt. Context windows up to 1M tokens (2M on Gemini 2.5 Pro). No extra setup required for multimodal — it just works.
Get Code Button
One click exports your entire prompt configuration — system instructions, examples, parameters, model selection — as working Python, JavaScript (Node.js), or REST curl commands. Copy directly into your project.
Deploy to Cloud Run
Single-button deployment wraps your prompt in a Cloud Run container and deploys it to Google Cloud. Fastest way to get a Gemini-powered API endpoint live without any infrastructure configuration.
Agentic Tools
Function calling (define tools, test how the model calls them), code execution (Gemini runs Python code in a sandbox), and Google Search grounding (responses are backed by live web search results). Test agentic patterns before implementing.
Pricing
AI Studio is free for individuals with a standard Google account — no credit card required. Rate limits apply at the free tier. For higher throughput, you connect an API key with billing enabled and pay per token used (standard Gemini API pricing applies).
AI Studio vs Vertex AI
Google's Gen AI SDK is designed to be endpoint-agnostic: code written against the Gemini Developer API in AI Studio runs on Vertex AI with a one-line change (swap the client initialisation and auth method; all prompt and model code stays identical). This means you prototype in AI Studio's friction-free environment and graduate to Vertex AI for production without rewriting anything.
Checklist
- Who is Google AI Studio designed for — end users or developers?
- What does the “Get code” button produce, and in which languages?
- What is the difference between the Gemini Developer API (used in AI Studio) and the Vertex AI API?
- Which agentic capabilities can you test directly within AI Studio?
- How do you move from a prototype built in AI Studio to a production deployment on Vertex AI?