Skip to main content

Google AI Studio: Vibe Coding Now for Subscribers, Apps from Prompts in Minutes

AI article illustration for ai-jarvis.eu
Google made vibe coding in AI Studio available to all Google AI Pro and Ultra subscribers starting April 20, 2026. Instead of writing code line by line, you describe the application in natural language — and Gemini builds the entire thing, including the backend. This change isn't just cosmetic: Google is trying to lure developers and tech-savvy non-pro creators away from tools like Cursor or GitHub Copilot.

What is vibe coding and why is everyone talking about it

The term vibe coding was first used by AI researcher Andrej Karpathy in early 2025. It's a method of application development where the programmer's main role shifts from writing code to directing an AI assistant — through conversation, refinement, and iteration. The result is a functional application, not just a code skeleton.

Karpathy's metaphor quickly caught on because it captures a real shift: today, a simple web application can be built in minutes, even without knowledge of React or Node.js. For experienced developers, it's a tool for rapid prototyping. For others, it's a gateway into the world of software.

What exactly Google launched on April 20, 2026

According to the official announcement on the Google blog, Google AI Pro and Ultra subscribers receive increased usage limits in AI Studio starting today. They now have direct access to Gemini Pro models within the application creation environment — without needing to enter a credit card or set up API keys.

Google frames this as a "low-setup billing bridge" — a transitional bridge for developers who want to try out a project before moving to a full-fledged API. Once the prototype is ready for deployment, you can switch to API keys directly within the AI Studio interface.

Practically, this means: if you pay for Google AI Pro or Ultra, open aistudio.google.com/apps, log in with your linked account, and start building.

Google AI subscription prices — what you get and for how much

Google offers three subscription tiers. At the lowest, Google AI Plus for 8 USD per month, you get 200 GB of cloud storage and access to the Gemini 3 Pro model. This tier does not include the new benefits in AI Studio.

Google AI Pro for 19.99 USD per month (2 TB storage, integration with Gmail and Google Docs) and Google AI Ultra for 249.99 USD per month (30 TB, access to Gemini 3.1 Pro, Veo 3.1 video generator, and the Mariner project for agency tasks) are the two tiers Google is targeting with the new AI Studio features.

For comparison: the competing Cursor costs 20 USD per month, GitHub Copilot approximately 10 USD per month. Google AI Pro for 19.99 USD now offers access to vibe coding as part of a broader package — without additional charge.

Antigravity: A billion-dollar editor bought by Google

Behind the vibe coding technology in AI Studio is Antigravity — an autonomous AI agent that plans, writes, tests, and debugs entire applications. Antigravity originated from the Windsurf team, which Google acquired for 2.4 billion dollars in 2025. It was launched as a standalone development environment in November 2025 and integrated into AI Studio on March 19, 2026.

Antigravity handles full-stack projects: frontend in React, Angular, or Next.js, backend via Firebase, database management, and third-party integrations. You control everything with natural language — the agent itself proposes architecture, fills in gaps, and suggests bug fixes.

Compared to the competition, Antigravity has one technological advantage: a parallel agent architecture that can work on multiple parts of a project simultaneously. In practice, this shortens the time from prompt to a functional application.

How Google stands against Cursor and GitHub

The market for AI development tools is crowded in 2026. Cursor has over 360,000 paying customers and is considered a professional standard with an emphasis on reliability and security (SOC 2 certification). Windsurf — whose team now works under Google — offered similar features at a lower price. And GitHub Copilot is deeply integrated into VS Code and the Microsoft ecosystem.

Google is betting on a different approach: AI Studio is a browser-based environment, without installation, connected to the Gemini API and the entire Google Cloud infrastructure. It primarily targets rapid prototyping and experimentation — not everyday production development in complex teams.

It's therefore for a different audience: not just professional developers, but also product managers, designers, or entrepreneurs who want to quickly test an idea.

Availability for Czechia and Europe

Google AI Studio is available in the Czech Republic and Slovakia — in English. Google AI Pro and Ultra subscriptions can be purchased via Google One in the EU, with prices in euros around 21.99 EUR per month for the Pro version (exact prices vary by country and current promotions). Vibe coding in AI Studio is available immediately after logging in with a linked Google One account.

For Czech developers or startups who already pay for Google Workspace or Google One Pro, this is added value without additional expense. Generated applications can be deployed on Google Antigravity hosting or on their own infrastructure — Google currently does not set strict limits for output formats.

From the perspective of the EU AI Act, code generation tools are classified as general AI systems with lower risk — they are not subject to stricter obligations like systems with a direct impact on people. This practically means that their deployment in European companies is not regulatorily complicated.

Practical perspective: when it makes sense and when it doesn't

Vibe coding in AI Studio makes the most sense for internal tools, prototypes, and MVP projects with lower load. If you need to quickly show a client how an application will work, or test a product before investing in full development, it's a significantly faster way than the classic approach.

Conversely — for production applications with high demands on security, compliance, or performance, Google itself recommends switching to API keys and a standard development pipeline. Antigravity is still a relatively new tool at this time, and larger, complex projects may encounter limitations.

Key figure: from prompt to a functional application can be achieved in a matter of minutes. This is a real change compared to the situation two years ago, when even a simple application meant hours of environment setup.

Do I need programming knowledge to use vibe coding in Google AI Studio?

Not necessarily. Vibe coding is designed to work even without deep code knowledge — describing the application in natural language is sufficient for basic projects. For more complex applications or debugging, a basic understanding of HTML, JavaScript, or databases is helpful, but not a prerequisite to start.

Are applications created in Google AI Studio owned by their creator, or does Google claim rights?

According to the Google One and Gemini API terms, user-generated outputs (including code) remain the property of the user. Google does not reserve rights to the generated code. However, for commercial use, it is always advisable to verify the current terms of service, especially when deploying in the EU.

How does Google AI Studio differ from tools like Cursor or GitHub Copilot?

Cursor and GitHub Copilot are primarily assistants for developers working with code — they help write, complete, and fix code. Google AI Studio with Antigravity goes further: instead of assisting in an editor, it autonomously designs and builds the entire application, including the backend. It's more comparable to Bolt.new or Lovable than to Cursor.