UIHarvest is a monorepo for scraping a live website and turning it into a structured design output (tokens, components, screenshots, assets, and AI design memory).

If you only want to run and test quickly, use apps/scraper.

• apps/scraper - main app for URL input, extraction, and output UI
• apps/server - backend for studio/chat/project APIs
• apps/web - studio frontend
• packages/* - shared auth/db/types/ui packages

• Bun 1.x
• Node.js 20+
• Chromium binaries for Playwright

From repo root:

bun install
bunx playwright install chromium

cp apps/scraper/.env.example apps/scraper/.env

Minimum required in apps/scraper/.env:

GOOGLE_CLOUD_API_KEY=your_key_here

Optional:

• SITE_PASSWORD (example/default: 654321)
• GOOGLE_CLOUD_PROJECT
• GCS_BUCKET

bun run --cwd apps/scraper dev

Open:

• UI: http://localhost:5173
• API: http://localhost:3333

In the UI:

1. Paste a website URL.
2. Start extraction.
3. Watch progress.
4. Review the final extracted output in the app.

After a successful run, UIHarvest gives you:

• Design tokens (colors, typography, spacing, radii, shadows)
• Component extraction with screenshots and metadata
• Downloaded assets (images, SVGs, fonts)
• Structured JSON output for programmatic use
• AI design memory files (when GOOGLE_CLOUD_API_KEY is set)

Typical output is saved under the scraper output folder and shown in the local UI while the job runs.

• Faster UI reverse-engineering from any live website
• Better handoff for developers and AI code generation
• Reusable design tokens for new projects
• Visual references (screenshots/assets) for accurate rebuilds
• Repeatable extraction flow for testing and benchmarking

From repo root:

bun run --cwd apps/scraper test
bun run --cwd apps/scraper typecheck

Manual smoke test:

1. Run bun run --cwd apps/scraper dev
2. Open http://localhost:5173
3. Enter https://example.com
4. Confirm extraction starts, progresses, and shows a final result

From repo root:

bun run dev
bun run build
bun run typecheck

gcloud auth login
gcloud config set project <your-project-id>
cd apps/scraper
./deploy.sh

• apps/scraper/README.md contains the full reproducibility guide used for submission.
• Without GOOGLE_CLOUD_API_KEY, AI phases are skipped.

MIT