How to export an AI-generated React app
Downloading source is the beginning of an export, not the end. A reliable handoff reproduces the runtime, explains dependencies, documents environment needs, and proves the project builds outside the generator.
The short answer
Freeze a known-good version, export source plus configuration, inspect the manifest, install in a clean directory, run type and production checks, test routes, document environment values, and commit the verified artifact before further work.
At a glance
Export bundle essentials
| Artifact | Purpose | Failure if missing |
|---|---|---|
| Source files | Application behavior and UI | The preview cannot be reproduced |
| package.json | Dependencies and scripts | Install and build are undefined |
| Build configuration | Framework and CSS pipeline | Local output differs from preview |
| Environment guide | External services and secrets | Runtime fails after deployment |
| Manifest and report | Traceability and known issues | Handoff begins without evidence |
1. Freeze the version you are exporting
Finish the acceptance pass before download and record the exact version or checkpoint. Do not export while an automated repair, streaming generation, or follow-up edit is still in progress.
2. Inventory the archive
Confirm the bundle includes every source file, entry point, package metadata, framework configuration, styles, public assets, shared UI support, environment guidance, and deployment files the preview relied on.
3. Read the manifest and diagnostics
A manifest should distinguish generated source from assembled support files and identify the intended package and app name. Diagnostics should report unresolved imports, protected paths, accessibility warnings, and bundle-level failures.
4. Install from a clean directory
Extract outside any existing monorepo so undeclared workspace dependencies cannot mask problems. Use the documented package manager, preserve its lockfile when provided, and note any peer-dependency or postinstall failures.
5. Run mechanical checks
Typecheck, lint, test where available, and create a production build. Start the built output and test direct navigation to nested routes, asset paths, environment failures, and browser refresh behavior.
6. Replace preview-only infrastructure
Identify builder-specific APIs, authentication, databases, storage, serverless functions, analytics, and environment values. Decide what moves with the project and what requires an external migration.
7. Create the first repository checkpoint
Once the clean export passes, initialize or connect the real repository and commit the untouched verified artifact. That baseline makes every subsequent product change reviewable.
FAQ
Frequently asked questions
Direct answers to the questions buyers and builders ask before committing a project to an AI app builder.
Is downloading a ZIP enough to own the code?
It gives you source possession, but practical ownership also requires the dependencies, configuration, environment information, and rights needed to build and operate it.
Why test in a clean directory?
Existing monorepo packages, global tools, caches, and environment variables can make an incomplete export appear healthy.
What does Squid include in a full export?
Squid currently assembles source and UI support files, package and build configuration, a README, manifest, quality and verification reports, and starter configs for Vercel, Netlify, and Cloudflare Workers.
Should I commit the export before editing it?
Yes. An untouched verified baseline makes later changes auditable and provides a reliable recovery point.
Keep researching
Related guides and comparisons
Build an app you can inspect, restore, and keep.
See the expected model cost before generation, review the resulting files, and export a verified React project when it is ready.