Skyscanner Flight Scraper (Apify Actor)

Scrapes Skyscanner flights via its internal web-unified-search JSON API — no browser in the hot path — over Apify's built-in Residential Proxy. Same technique as jupri/skyscanner-flight.

input → resolve IATA→entityId → mint cookie → POST web-unified-search → poll until complete → parse → dataset

Why no browser

Skyscanner's web-unified-search endpoint returns pure JSON. The only obstacle is the PerimeterX cookie (_px3). A clean residential IP usually gets a usable cookie with a plain HTTP GET (Phase 1). If that IP is challenged, the actor falls back to a headless anti-detect browser (Camoufox, Phase 2) that loads one page from the same sticky IP and hands back the cookie. Everything else is HTTP.

Input

Mirrors jupri's fields. Minimum: origin.0, target.0, depart.0.

Field	Meaning
origin.0..5 / target.0..5 / depart.0..5	Per-leg origin, destination, and YYYY-MM-DD date. For a round trip: leg 0 = outbound, leg 1 = origin.1=target.0, target.1=origin.0, depart.1=return.
market / currency	ISO country + currency (e.g. IL / USD). currency is required internally.
sort	score | cheapest | fastest | departure.
cabin_class	economy | premiumeconomy | business | first.
adults / children / infants	Passenger counts (max 8 each).
non_stop / one_stop / two_stop	Stop filters (any combination).
limit	Slice the result count.
dev_proxy_config	Proxy. Defaults to Apify RESIDENTIAL (needed to pass PerimeterX).
dev_custom_headers / dev_custom_cookies	Extra headers / cookies (inject a pre-minted _px3 here).
dev_transform_fields / dev_no_strip / dev_dataset_name / dev_dataset_clear	Output shaping.

Example:

{
  "origin.0": "TLV", "target.0": "HER", "depart.0": "2026-08-02",
  "origin.1": "HER", "target.1": "TLV", "depart.1": "2026-08-05",
  "market": "IL", "currency": "USD",
  "cabin_class": "economy", "adults": 1, "non_stop": true,
  "sort": "cheapest"
}

Output

One dataset record per itinerary:

{
  "id": "16995-2608021920--30967-0-12130-2608022110|12130-2608051125--30967-0-16995-2608051305",
  "price": 229, "currency": "USD",
  "carriers": "Blue Bird Airways",
  "stops": 0,
  "origin": "TLV", "destination": "HER",
  "departure": "2026-08-02T19:20", "arrival": "2026-08-02T21:10",
  "durationMinutes": 110,
  "legs": [ … ], "agents": [ … ]
}

id is the legOut|legIn config id — the key for a per-agent price lookup (/g/sonar/v3/itinerary/details/poll).

Run

Locally (needs Apify CLI + an APIFY token for proxy):

$apify run --purge

Deploy:

$apify push

Call it over the API (this is how the .NET admin consumes it):

curl -X POST \
  "https://api.apify.com/v2/acts/<actor-id>/run-sync-get-dataset-items?token=<APIFY_TOKEN>" \
  -H 'Content-Type: application/json' \
  -d '{"origin.0":"TLV","target.0":"HER","depart.0":"2026-08-02","market":"IL","currency":"USD"}'

Gotchas baked in

• user-agent on the search POST must match the UA the cookie was minted with (Phase 2 returns the browser's exact UA for this reason).
• x-skyscanner-currency is required (missing → 400).
• Mint + every fire exit the same sticky IP (PerimeterX binds _px3 to the IP).
• First response ≈ 10 results (status: incomplete); it grows toward ~860 as the actor polls to complete.