Google's AI Studio Build Now Syncs GitHub Repos Both Ways With Gemini

Google just closed a workflow gap that has been quietly frustrating anyone using AI Studio Build. The vibe-coding environment can now import existing GitHub repositories with one click and keep them synced in both directions, meaning code you write locally and code Gemini writes in the browser finally live in the same place.

The one-way street is gone

Until this update, AI Studio Build only shipped code outward. The relationship between AI Studio Build and GitHub only ran in one direction. Developers could export their finished work out to a GitHub repository, but there was no way to bring an existing codebase back in. That limitation is gone.

The new entry point is a button labeled Import from GitHub, tucked under the + icon in the prompt input box. Point it at a repo and the code loads directly into the Build environment, ready for Gemini to start editing.

How the sync actually works

Once a repo is linked, changes flow both ways automatically. You can push changes prompted in AI Studio directly to your repository with AI-generated commit messages, or pull changes made locally in your IDE or by teammates back into AI Studio. The sync status is managed from a dedicated GitHub tab under Settings.

Merge conflicts are handled inside the Studio too. When AI Studio and GitHub disagree on a file, a Resolve conflicts dialog shows a side-by-side diff and lets you pick which version wins per file. That's a small detail but a meaningful one, because it means you don't have to bail out to the command line the moment two branches diverge.

One thing worth knowing: imported code isn't just displayed as-is. The more consequential piece of the update is the sync mechanism itself. Once a repository is connected, changes made inside AI Studio Build flow back out to GitHub automatically, and any updates pushed to the GitHub repository show up inside the Studio. Behind the scenes, the platform reshapes imported code to run inside its own runtime, so your project is being adapted to Google's environment rather than just mirrored.

What you can actually build with it

AI Studio Build isn't a general-purpose IDE. It's opinionated about the stack. When you create a new app that uses the Gemini API, AI Studio automatically configures your Gemini API key as a secret in the app's server-side environment. You can view and manage this key in the Secrets panel. Automatic setup: Your GEMINI_API_KEY is set up for you , no manual configuration is required to start building. Server-side only: API keys are injected into the server-side runtime and are never included in client-side code.

For web apps, the default frontend is React with a Node.js server-side runtime that supports npm packages, Firebase auth and Firestore, and Google Workspace API integrations. There's also a native Android track using Kotlin and Jetpack Compose. Python-heavy repos will get you further than raw C++, but the sweet spot is clearly modern JavaScript stacks.

Who this actually helps

The immediate winners are people carrying around half-finished side projects or older codebases they don't want to rewrite from scratch. A few concrete cases:

• A hackathon repo you never finished, now importable so Gemini can help patch it up
• A legacy app that needs incremental modernization instead of a full rewrite
• A teammate onboarding onto an unfamiliar codebase who wants chat-based Q&A over the files
• A prototype started in AI Studio that outgrew the browser and needs to move to a real IDE, then come back for quick UI iterations

For teams managing older codebases, the update offers a way to modernize gradually instead of committing to a full rewrite. Developers can import the existing code, apply AI-assisted improvements piece by piece, and push those changes back to GitHub, all without leaving the Studio environment. That incremental approach lowers the risk that typically comes with large-scale rewrites, where something breaks and nobody can pinpoint exactly why.

Where it will probably struggle

The honest question is how the sync holds up on real projects. Anything that depends on native binaries, unusual build systems, or non-standard runtimes is likely to trip over the code normalization step. And because AI Studio wraps everything in its own server-side container, repos that expect specific infrastructure won't map cleanly. Expect the smoothest experience with vanilla React, Next.js, and Python projects that already fit the shape Google expects.

Getting started

The feature is live in AI Studio Build at no cost beyond standard Gemini API usage. Finished projects can be pushed to a GitHub repo for collaboration or deployed to Cloud Run, where standard Cloud Run pricing applies. If you're already living in the Gemini ecosystem, this removes one of the last real reasons to avoid it for anything larger than a demo.