Broad public API directory with a strong AI and agent focus.
V1 is intentionally simple:
- static, build-time rendered pages
- Cloudflare Worker + static assets
- free APIs only
- multiple public GitHub sources
- strong category and capability landing pages
- lightweight client-side search on top of a crawlable archive
- TypeScript for the Worker and build pipeline
- Tailwind CSS v4
- vanilla browser JavaScript for search/filter behavior
- Wrangler for Cloudflare deployment
public-api-lists/public-api-listspublic-apis/public-apistools-collection/apis-collection
npm run buildbuilds the full static site intodist/npm run typecheckruns TypeScript checksnpm run checkruns type checks pluswrangler types --checknpm run cf:typegenregeneratesworker-configuration.d.tsnpm run devbuilds the site, then starts Wrangler local developmentnpm run audit:linksaudits all target docs URLs with 10 parallel workers and refreshes the ignore listnpm run screenshots:capturecaptures real screenshots into.cache/screenshotsusing Playwright
V1 supports two modes:
- default: generated SVG placeholders for every API entry
- local/CI capture mode:
npm run screenshots:capturerenders real screenshots with Playwright into.cache/screenshots/ - optional remote mode: if
SCREENSHOT_API_BASE_URLis provided, the build tries to fetch real screenshots and caches them in.cache/screenshots/
If remote screenshot capture is unavailable, the build does not fail.
Recommended production path:
- run
npm run audit:linksfirst, so dead targets land indata/api-ignore-list.json - run
npm run screenshots:capture -- --concurrency=10 - then run
npm run build
The deploy workflow restores and saves .cache/screenshots, installs Chromium, refreshes missing screenshots plus screenshots older than 90 days, and then deploys.
The repository includes GitHub Actions workflows for CI and Cloudflare deployment.
Required secrets for deploy:
CLOUDFLARE_API_TOKENCLOUDFLARE_ACCOUNT_ID
Optional screenshot-related secrets:
SCREENSHOT_API_BASE_URLSCREENSHOT_API_TOKEN