Imagine getting free access to some of the world’s most advanced AI models for text, images, audio, and even video - all in one place. No installation, and no subscription cost. That’s Google AI Studio, and in this article, I’ll give you a complete tour of everything you can build, prompt, and test. I’ve been recommending this to NZ businesses as a great starting point for experimenting with AI.
Getting Started
Search for Google AI Studio and sign in with a free Google account. You’ll see a chat interface similar to ChatGPT or Gemini, but with extensive customisation options on the side panel.
You can select from all of Google’s released models -2.5 versions, older 2.0 and 1.5 models, open-source Gemma versions, and even Learn LM for tutoring. For this tour, we’ll focus on Gemini 2.5 Pro.
Chat Window Features
Beyond basic chat, Google AI Studio offers powerful extras:
- Structured output: JSON format for use in your code
- Code execution: Upload CSV data and execute Python code directly - no hallucination on data analysis
- Function calling: Connect custom APIs
- Google Search and URL context: Ground responses with live web data
- Temperature and Top P controls: Fine-tune how the model responds
It’s fully multimodal - upload video, audio, text, images, files from Google Drive, and even YouTube videos. All completely free with no limits.
Real-Time Streaming
The stream section lets you talk to Gemini in real time. You can speak directly, share your webcam, or share your screen. This is perfect for walking through exams, getting help with complex software, or having the AI teach you how to use unfamiliar tools while you’re on them.
Image Generation with Imagen
Under Generate Media, you’ll find Imagen for image generation. You can set aspect ratios, number of results, and choose between models. Getting four results per prompt gives you a much better chance of getting exactly what you’re looking for - compared to other tools that return just one.
Speech Generation
There are two modes: single speaker and multi-speaker audio. Generate voiceovers, podcasts, or conversations. Choose between Flash (fast) and Pro (detailed) models, and customise voices and tones. You can download and export everything.
Music Generation with Lyria
Lyria is Google’s music generation model, available as a Prompt DJ. Adjust knobs for different music styles and watch it blend genres together in real time. The API is available through Firebase Studio for generating music programmatically.
Video Generation with Veo
Veo is the video generation model. Choose aspect ratios, set negative prompts to exclude unwanted elements, and even upload a reference image. The upcoming Veo 3 model adds built-in audio generation with dramatically improved quality.
Build Full Web Applications
The Build section works like Lovable, Cursor, or Bolt - create full web applications from a single prompt using Gemini 2.5 Pro. I built a dinosaur jumping game in 89 seconds from one small prompt. You can iterate on graphics, features, and design, then deploy directly to Google Cloud.
Final Thoughts
Google AI Studio is arguably the most important AI tool to learn right now because it’s completely free and offers text, image, video, and audio generation with the latest models - all in one place.
