Skyscanner Flight Tracker

A weekend flight-deal platform that beats Skyscanner's PerimeterX anti-bot — mobile-API PX cookie solving + residential proxies feeding a self-hosted deal tracker with price history and 200+ destination coverage.

🖥 live demo🔒 private repo

Runs privately — walkthrough available on request.

The Build Loop

Because I hunt weekend flight deals from Indianapolis, I kept hitting the fact that Skyscanner blocks every scraper behind PerimeterX — instant 403s, JS challenges, fingerprinting, so I built a two-part system: a scraper that solves PX challenges on Skyscanner's mobile API through residential proxies, and a self-hosted tracker app that turns the feed into price history and deal alerts, that replaces "refresh Skyscanner for an hour" with a dashboard that watches 200+ destinations daily. It taught me that anti-bot bypass at this level is a systems problem — cookie solving, proxy rotation, and API choice matter more than any single trick.

What it looks like

Screens and recordings for this project are being added.

Skyscanner Flight Tracker

One-liner: A weekend flight-deal platform that beats Skyscanner’s PerimeterX anti-bot — mobile-API PX cookie solving + residential proxies feeding a self-hosted deal tracker with price history and 200+ destination coverage.

Origin (the Build Loop)

Because I hunt weekend flight deals from Indianapolis, I kept hitting the fact that Skyscanner blocks every scraper behind PerimeterX — instant 403s, JS challenges, fingerprinting, so I built a two-part system: a scraper that solves PX challenges on Skyscanner’s mobile API through residential proxies, and a self-hosted tracker app that turns the feed into price history and deal alerts, that replaces “refresh Skyscanner for an hour” with a dashboard that watches 200+ destinations daily. It taught me that anti-bot bypass at this level is a systems problem — cookie solving, proxy rotation, and API choice matter more than any single trick.

The problem

Skyscanner sits behind PerimeterX: plain requests get fingerprinted and blocked, headless browsers get challenged, and there’s no public API for the flexible-date searches that actually find weekend deals. The only paths in are the mobile API (with PX cookie solving) or a fully stealthed real browser.

What it does

  • Scraper — three-tier fallback: Skyscanner mobile API (PX cookie solver + residential proxy) → Playwright stealth browser → Bright Data managed proxy.
  • Tracker app (SkyApp) — self-hosted FastAPI app: set a home airport, track cheapest weekend trips to major destinations, background refresh (on-demand + daily scheduled), price history, visual comparison, one-click booking links.
  • 200+ destinations with dynamic airport mapping, served from a VPS API on :8010.

How it’s built

  • Stack: Python (both sides), Playwright (stealth-patched), PX cookie solver, residential proxies (DataImpulse), FastAPI + APScheduler + SQLite (app), VPS API + Replit UI.
  • Notable engineering:
    • PerimeterX bypass on the mobile API — solved PX cookies + residential IP rotation, chosen over browser automation for speed; browser path kept as fallback.
    • Two-repo architecture — the scraper (credential-heavy, VPS) is decoupled from the tracker (public-facing, Replit), so proxy credentials never touch the web app.
    • APScheduler background refresh — daily sweep + on-demand polls, SQLite price history underneath.
  • Architecture: cron/daily trigger → mobile-API search (PX solver + proxy) → price feed → VPS API :8010 → SkyApp dashboard (Replit) with price history + booking links.

Proof points

  • Working PX bypass against one of the most aggressive anti-bot deployments in travel.
  • 200+ destinations tracked daily with price history and visual comparison.
  • Running in production — VPS API auto-starts on boot; the app refreshes on a schedule without babysitting.

What to show

  • Demo: The SkyApp dashboard (weekend grid + price history). The scraper side is credential-heavy — case-study-only.
  • Visuals needed: the weekend-deal grid for a home airport; a price-history chart for one route; the architecture diagram (mobile API → PX solver → proxy → feed → app).

Case-study angle

The portfolio’s deepest anti-bot work: a PerimeterX bypass good enough to run daily against Skyscanner, wrapped in a two-repo architecture that keeps credentials off the public app. Scraper-as-infrastructure, not scraper-as-script.