Google Flights Scraper

Get live Google Flights prices as flat, ready-to-use rows - one row per itinerary, with airlines, times, duration, stops, layovers, aircraft, and CO2 estimates. One-way or round-trip, any cabin class, any currency. No charge for failed fetches.

Try it

Paste this into the input editor and hit Run:

{
  "origin": "LHR",
  "destination": "JFK",
  "departDate": "2026-09-25",
  "returnDate": "",
  "adults": 1,
  "cabinClass": "economy",
  "currency": "GBP",
  "maxItems": 50
}

Input fields

Field	Type	Values	Notes
origin	string	3-letter IATA code, e.g. LHR, JFK, NYC	Airport codes and metro/city codes both work
destination	string	3-letter IATA code	Same as origin
departDate	string	YYYY-MM-DD	Must be today or later
returnDate	string	YYYY-MM-DD or empty	Empty = one-way. When set, rows are departing-flight options priced as round-trip totals (how Google Flights itself works)
adults	integer	1-9	Default 1
cabinClass	string	economy | premium_economy | business | first	Default economy
currency	string	ISO code, e.g. USD, GBP, EUR	Prices returned in this currency
maxItems	integer	e.g. 50	Caps rows delivered before the run stops
proxyGroups	array	e.g. ["RESIDENTIAL"]	RESIDENTIAL (default) recommended - Google blocks datacenter IPs on the Flights surface

One query returns the complete result set for that route and date (typically 10-30 itineraries) in a single request - there is no pagination to pay for.

Output

Flat rows, one per itinerary. No nested objects to unpack - multi-leg details come as compact |-delimited strings. Real example row (LHR-JFK one-way, economy):

{
  "origin": "LHR",
  "destination": "JFK",
  "departDate": "2026-09-25",
  "returnDate": "",
  "cabinClass": "economy",
  "adults": 1,
  "currency": "GBP",
  "fetchedAt": "2026-08-26T12:00:00.000Z",
  "category": "best",
  "price": 602,
  "priceCurrency": "GBP",
  "airlineCodes": "B6",
  "airlineNames": "JetBlue",
  "departureAirport": "LHR",
  "departureTime": "2026-09-25 08:15",
  "arrivalAirport": "JFK",
  "arrivalTime": "2026-09-25 11:19",
  "durationMinutes": 484,
  "stops": 0,
  "layovers": "",
  "flightNumbers": "B62220",
  "legs": "B62220 LHR 2026-09-25 08:15 -> JFK 2026-09-25 11:19",
  "aircraft": "Airbus A321neo",
  "co2Grams": 537000,
  "co2TypicalGrams": 484000,
  "deepLink": "https://www.google.com/travel/flights/search?tfs=GhoSCjIwMjYtMDktMjVqBRIDTEhScgUSA0pGS0IBAUgBmAEC&hl=en&curr=GBP"
}

A connecting itinerary fills the multi-leg fields:

{
  "airlineNames": "Tap Air Portugal",
  "price": 519,
  "stops": 1,
  "layovers": "LIS 490m",
  "flightNumbers": "TP1363 | TP209",
  "legs": "TP1363 LHR 2026-09-25 06:00 -> LIS 2026-09-25 08:45 | TP209 LIS 2026-09-25 16:55 -> JFK 2026-09-25 20:05",
  "durationMinutes": 1145
}

• category is best (Google's ranked picks) or other (the rest of the result set).
• deepLink opens the exact same search in Google Flights - hand it to users for booking.
• co2Grams is Google's per-passenger CO2e estimate for the itinerary; co2TypicalGrams is the route-typical value it is compared against.

Queries that fail after retries are logged to a separate google-flights-errors dataset (origin, destination, departDate, reason, fetchedAt) instead of leaving a silent gap in your results.

Coming from fast-flights?

The fast-flights Python library proves this route works - but you run it, patch it when Google changes internals, and supply your own unblocked IPs. This actor is the same server-rendered ?tfs= route as a managed service: residential proxy rotation, session-scoped cookies, retries with identity rotation, block detection, and flat rows with layovers and flight numbers already extracted.

fast-flights	This actor
FlightData(date=..., from_airport="LHR", to_airport="JFK")	"origin": "LHR", "destination": "JFK", "departDate": ...
Passengers(adults=2)	"adults": 2
seat="business"	"cabinClass": "business"
trip="round-trip" + second FlightData	just set "returnDate"
bring your own proxy	residential rotation built in
Result.flights[n].price	flat rows with price, layovers, flight numbers, CO2

How this compares

	This actor	Typical Flights actors on the Store
Charge on failed queries	Never - only delivered rows are billed	Often bill per run or per started query
Actor start	Free - no per-run fee	Some charge a flat fee before any data lands
Output shape	Flat rows, one per itinerary	Frequently nested per-search JSON you have to unpack
Reliability approach	Session rotation (Crawlee session pool), bounded retries with identity rotation, block detection	Not usually documented in the listing

Pricing

Pay-per-event, priced on delivered rows: you're billed only once itinerary rows land in your dataset. Starting a run costs nothing. If a query fails after retries, it goes to the errors dataset instead, and you pay nothing for it.

Using the API

Apify API (HTTP)

curl "https://api.apify.com/v2/acts/meticulous_ground~google-flights-scraper/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "origin": "LHR",
    "destination": "JFK",
    "departDate": "2026-09-25",
    "currency": "GBP"
  }'

JavaScript (apify-client)

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });

const run = await client.actor('meticulous_ground/google-flights-scraper').call({
  origin: 'LHR',
  destination: 'JFK',
  departDate: '2026-09-25',
  returnDate: '2026-10-02',
  cabinClass: 'economy',
  currency: 'GBP',
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

Python (apify-client)

import os
from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])

run = client.actor("meticulous_ground/google-flights-scraper").call(run_input={
    "origin": "EDI",
    "destination": "LHR",
    "departDate": "2026-09-25",
    "currency": "GBP",
})

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item)

Run one query per Actor call - each call returns that route and date's complete itinerary set.