Academy
5. Reviewing and Exporting Your Code
Inspect the generated files inside the workspace and download them so you can continue development in your own environment.
Once the AI has generated and iterated on your project, every file it produced lives inside the Code Builder workspace. You can inspect the full file tree — source code, assets, configuration — directly in the interface, right alongside the live sandbox preview. This is not a black box: the generated files belong to you, and the workspace gives you read access to review exactly what the AI wrote before you decide to export anything. Understanding the structure here helps you evaluate code quality, spot issues, and plan your next moves outside the platform.
The live preview running beside the conversation is your real-time validation tool. Every iteration you requested refreshed this preview, so by the time you are ready to export, what you see in the preview is exactly what the downloaded code will produce when you run it locally. The preview runs inside an isolated cloud sandbox — completely separate from your machine and other users — so the behavior you observe is a clean, reproducible execution of the generated code, not influenced by any local environment quirks.
Before downloading, do a final review pass in the workspace. Check that the execution profile matches your local target environment (runtime, capabilities). If something looks off, iterate one more time conversationally — it costs only a few AI credits and is far easier than patching files after export.
Exporting is straightforward: download the generated files from the workspace. You get the complete project — all source files the AI created or modified across your iterations. From there, you can open the project in your own IDE, wire it into your CI/CD pipeline, add it to version control, or deploy it however you like. The code is yours with no lock-in; the platform is a starting accelerator, not a cage.
After downloading, compare the execution profile you used in Code Builder with your local dev environment. If the sandbox runtime differs from your local stack, adjust dependencies or configurations accordingly to ensure the app behaves identically outside the platform.
Open your Code Builder workspace at the project route. Browse the generated file tree and inspect at least two files (e.g., the main entry point and a component). Verify the live preview still matches your expectations, then download the full project. Open it in your local IDE and confirm it runs.
The Code Builder workspace gives you full visibility into every file the AI generated. The live sandbox preview guarantees what you see is what you export. Downloading hands you a complete, portable project — no lock-in. Before exporting, review the code in-workspace, verify the preview, and check that your execution profile aligns with your local environment. After download, your standard dev workflow takes over.