toon_stream_api — got-scraping edition

What changed from the original

Original	This version
src/baseurl.txt fetched from GitHub on every cold start	src/config.js — hardcoded constant, edit once
src/cf_proxy.txt fetched from GitHub, proxied every request	Removed entirely
lib/scrape.js used Cloudflare Workers cache + CF proxy fallback	Direct got-scraping requests (browser TLS fingerprinting)
api/category.js, api/scrape.js, api/extra.js each had their own inline CF proxy fetch	Now all use fetchPage() from lib/scrape.js

All API routes, response shapes, and scraping logic are identical.

---

Changing the target URL

Open src/config.js and update BASE_URL:

export const BASE_URL = 'http://toonstream.dad'; // ← change this

That's it. No other files need touching.

---

Setup

npm install
vercel dev        # local dev
vercel            # deploy

---

API Routes

Route	Description
GET /	Site meta + homepage article
GET /home	Home page: featured, episodes, series, movies, schedule
GET /search/:query	Search results
GET /series/:slug	Series details + seasons/episodes
GET /episode/:slug	Episode detail + video servers
GET /movies/:path	Movie detail + video options
GET /movies	Movie listing (paginated)
GET /series	Series listing (paginated)
GET /letter/:letter	Browse by letter
GET /cast/:cast	Cast page
GET /category/:path	Category page
GET /api/catalog?type=series|movies&page=N	Full catalog
GET /api/scrape?section=menu|footer|schedule|4|5	Structural data
GET /api/extra?section=...	Extended structural data
GET /api/embed?url=...	Resolve embed iframe URL
GET /api/s_movies?url=...	Scrape any movie list URL

---

Why got-scraping?

got-scraping rotates real browser TLS fingerprints (JA3/JA4) on every request, making the scraper much harder to block than a plain fetch() with a spoofed User-Agent. It is a drop-in async HTTP client — no Cloudflare proxy needed.