Which Google Flights scraper should you use? Ten options compared
Last updated
Quick answer: there is no official Google Flights API, so every option reads the public results page. For proven volume at a low rate, memo23 at $0.001 per result plus $0.005 per run start (476 users, 99.9% success) is the value pick, and johnvc is the most used at 2,480 users with a 4.0 rating. Datapika's Google Flights scraper at $0.0015 per itinerary ($1.50 per 1,000, 3,696 runs, 93.5% success) fits when you need CO2 in grams, 30 or more result languages, multi-city trips and a hard per-query row cap. Skip the $19.99 a month rental at 0% success.
What happened to the official Google Flights API?
There is no Google Flights API you can sign up for. Google's only public airfare feed was QPX Express, built on the ITA Software technology it bought for $700 million in a deal that closed in 2011. On November 1, 2017 Google told TechCrunch it would retire the API "given the low interest among our travel partners", and it switched the service off on April 10, 2018. Before the shutdown the API cost $0.02 per query after 50 free queries a month, down from an earlier $0.035. The enterprise QPX product continued for large contracted travel companies, and a Google spokesperson said it generated the vast majority of the travel data revenue. Eight years later nothing public has replaced it, so every option on this page reads the Google Flights results page and converts what a traveler sees into rows. The tools differ in what they bill (per itinerary, per search, per results page, per month), how much of the page they return, and how often a run actually finishes.
- QPX Express retirement announced November 1, 2017; shut down April 10, 2018 (TechCrunch)
- Google's stated reason: low interest among travel partners
- Last public price was $0.02 per query after 50 free queries a month
- No public replacement as of August 29, 2026; enterprise QPX is contract-only
- Every scraper below parses the public results page, so success rate is what separates them
How do the best Google Flights scrapers compare on price, rating and users?
Ten actors were checked on the Apify Store on August 29, 2026, and only three carry reviews: johnvc's Google Flights API at 4.0 stars from 6 reviews, memo23 at 5.0 from a single review, and makework36 at 1.0 from one review. The rest are unrated, so the table leans on the two numbers Apify publishes for every actor, total users and 30-day success rate. Johnvc has by far the largest base at 2,480 users and 377 in the last 30 days, but bills $0.03 per results page plus $0.01 per run on the free plan, the most expensive per search here. Memo23 is the value pick with volume behind it: $0.001 per result plus $0.005 per start, 476 users and 99.9% success. Automation-lab is cheaper still at $0.00023 per result on the free plan plus $0.005 per start, with 396 users and 100%. Kaix lists $0.06 per 1,000 on the free plan but bills Apify compute on top. Datapika charges $0.0015 per itinerary plus platform usage, with 3,696 lifetime runs and 93.5% success. Scrape.badger bills $0.012 per search, skootle $8 per 1,000 on the free plan, and ScrapeBase rents at $19.99 a month with 0% of runs succeeding.
- Only rated actors: johnvc 4.0 (6 reviews), memo23 5.0 (1 review), makework36 1.0 (1 review)
- Largest user base: johnvc, 2,480 users and 377 in the last 30 days
- Cheapest listed rate with no compute bill: automation-lab at $0.23 per 1,000 on the free plan plus $0.005 per start
- Highest success among actors with real volume: automation-lab and johnvc 100%, memo23 99.9%, Datapika 93.5%
- Only per-search billing: scrape.badger at $0.012 per search, documented as an MCP tool
- Only monthly rental: ScrapeBase at $19.99 plus usage, 0 monthly users and 0% success
What does it cost to scrape 1,000 and 10,000 Google Flights itineraries?
Figures use each actor's free-plan rate on August 29, 2026 and assume 25 itineraries per search, which is Datapika's default max_results and inside the 30 to 80 Google usually shows, so 1,000 itineraries is 40 searches. Datapika: 1,000 times $0.0015 is $1.50 and 10,000 is $15.00, plus Apify platform usage for the run time, which the listing does not fix in advance. Memo23: $1.00 in results plus 40 starts at $0.005 is $1.20; 10,000 is $10.00 plus $2.00, $12.00. Automation-lab: $0.23 plus $0.20 in starts is $0.43; 10,000 is $2.30 plus $2.00, $4.30. Scrape.badger: 40 searches at $0.012 is $0.48 and 10,000 is $4.80, but if you keep only the five cheapest fares per route the same 1,000 rows need 200 searches and cost $2.40. Johnvc at one results page per search: 40 pages at $0.03 plus 40 runs at $0.01 is $1.60, and 10,000 is $16.00. Skootle: $8.00 plus $0.20 is $8.20, and $82.00 for 10,000. Apify's own $0.00005 start and $0.00001 dataset-item accounting events add under 2 cents per 1,000 rows on any of them.
- 1,000 itineraries: automation-lab $0.43, scrape.badger $0.48, memo23 $1.20, Datapika $1.50 plus compute, johnvc $1.60, skootle $8.20
- 10,000 itineraries: automation-lab $4.30, scrape.badger $4.80, memo23 $12.00, Datapika $15.00 plus compute, johnvc $16.00, skootle $82.00
- Per-search billing flips against you when you cap results low: 5 rows per search makes scrape.badger $2.40 per 1,000
- Kaix is left out of the arithmetic because its compute charge is not published on the listing
- Paid Apify plans cut most rates: automation-lab drops to $0.12 per 1,000 and skootle to $5.00 on Gold
Which fields and features differ between Google Flights scrapers?
Price and stops are universal; the differences are in trip types, locale, carbon data and agent readiness. Multi-city itineraries are supported by Datapika, johnvc, memo23, kaix, scrape.badger and ScrapeBase; automation-lab and Scrape Sage list one-way and round-trip only. Datapika accepts a two-letter ISO code and returns airline and airport names in 30 or more result languages, johnvc lists 29 or more languages and 39 countries, memo23 takes language, currency and market inputs, and automation-lab lists multi-language and multi-currency output. Carbon data varies: Datapika and skootle report grams for the itinerary plus Google's typical figure for the route, automation-lab, scrape.badger and ScrapeBase report kilograms or a percentage versus typical, and johnvc and memo23 document no CO2 field. Skootle adds a price_insights row with median, P10 and P90 fares, a 0 to 100 completeness score and an agentMarkdown field. Kaix keeps the raw Google response next to normalized rows and adds calendar prices. Booking links come from johnvc as a $0.02 per option add-on and from makework36, which compares Google Flights with Kiwi, Travelpayouts, Ryanair, easyJet, Wizz Air and Norwegian. Datapika echoes the query inputs and a scraped_at timestamp on every row so fan-out runs merge by route and date.
- Multi-city: Datapika, johnvc, memo23, kaix, scrape.badger, ScrapeBase; not automation-lab or Scrape Sage
- Languages: Datapika 30+ via ISO code; johnvc 29+ languages and 39 countries; memo23 language, currency and market
- CO2: Datapika and skootle in grams with route typical; automation-lab, scrape.badger, ScrapeBase in kg or percent; johnvc and memo23 none documented
- Booking links: johnvc as a paid add-on, makework36 across seven sources; Datapika returns none
- Agent extras: skootle agentMarkdown and completeness score; Datapika echoed inputs plus scraped_at on every row
When should you choose which Google Flights scraper?
Pick by the shape of your workload, not the headline rate. If you want the most battle-tested option and can pay about 4 cents per search, johnvc's 2,480 users, 4.0 rating and 100% success on August 29, 2026 make it the low-risk default, and its booking-link add-on is unique among the Google-only actors. If you pull whole result pages in bulk, memo23 at $1.00 per 1,000 plus $0.005 per start with 99.9% success, or automation-lab at $0.23 per 1,000 plus the same start fee, are the cheapest options with real volume behind them. If your agent asks one question at a time and wants an MCP tool that returns the full page, scrape.badger's $0.012 per search is simple to reason about. Choose Datapika at $1.50 per 1,000 when you need CO2 in grams, 30 or more result languages, multi-city trips and a max_results cap of 1 to 200 that bounds itinerary charges before a run starts; accept that platform usage is billed on top and that its 93.5% success trails the leaders. Choose skootle only for its fare percentiles, and skip ScrapeBase until its 0% success rate changes.
- Most used, rated, booking links: johnvc, at the highest cost per search
- Bulk fare history at the lowest proven cost: memo23 or automation-lab
- Agent tool with per-search billing: scrape.badger
- CO2 in grams, 30+ languages, multi-city, per-query row cap: Datapika
- Fare percentiles and LLM-ready markdown: skootle, at a premium and 59.3% success
- Cross-site price comparison with booking links: makework36, despite its 1.0 rating
| Tool | Price (free plan) | Rating (reviews) | Users (30-day) | Success rate | Pricing model | Best for |
|---|---|---|---|---|---|---|
| Datapika google-flights-scraper (openclawai) | $1.50 per 1,000 itineraries + Apify platform usage | No reviews yet | 3,696 lifetime runs | 93.5% | Per itinerary plus platform usage | CO2 in grams, 30+ languages, multi-city, 1 to 200 row cap, MCP |
| johnvc/google-flights-data-scraper-flight-and-price-search | $0.03 per results page + $0.01 per run | 4.0 (6) | 2,480 (377) | 100% | Per page plus setup fee, no compute | Largest base, booking-link add-on, 29+ languages |
| memo23/google-flights-scraper | $1.00 per 1,000 results + $0.005 per start | 5.0 (1) | 476 (199) | 99.9% | Per result plus start fee, no compute | Cheapest proven bulk pulls, multi-city |
| automation-lab/google-flights-scraper | $0.23 per 1,000 + $0.005 per start ($0.12 on Gold) | No reviews yet | 396 (66) | 100% | Per result plus start fee, no compute | Lowest per-row rate, up to 200 results per search |
| kaix/google-flights-scraper | $0.06 per 1,000 + Apify compute ($0.05 on Gold) | No reviews yet | 372 (181) | 92.8% | Per result plus platform usage | Raw Google responses, calendar prices |
| scrape.badger/google-flights-scraper | $0.012 per search ($12 per 1,000 searches) | No reviews yet | 10 (1) | 98.2% | Per search, no compute | Agents calling one route at a time via MCP |
| skootle/google-flights-scraper | $8.00 per 1,000 + $0.005 per start ($5.00 on Gold) | No reviews yet | 38 (4) | 59.3% | Per record plus start fee, tiered | Fare percentiles, agentMarkdown, completeness score |
| makework36/flight-price-scraper | $2.50 per 1,000 results | 1.0 (1) | 873 (101) | 100% | Per result, no compute | Cross-site comparison with booking links |
How to do it
- 1.Open https://apify.com/openclawai/google-flights-scraper, enter origin and destination IATA codes such as JFK and LAX, a departure date, and optionally a return date or multi-city legs.
- 2.Set trip type, cabin, passenger mix, a two-letter language code and a max_results cap between 1 and 200; at $0.0015 per itinerary the cap bounds the itinerary charges for that query, with Apify platform usage billed on top.
- 3.Run it and export JSON, CSV or Excel, or call the run-sync-get-dataset-items endpoint from Python or JavaScript with your Apify token to get the itinerary array in one request.
- 4.For AI agents, add the actor as a tool through https://mcp.apify.com/?tools=fetch-actor-details,openclawai/google-flights-scraper so a Claude or GPT assistant can request fares directly.
Questions, answered
Is there an official Google Flights API I should use instead?
No. Google's QPX Express API, built on the $700 million ITA Software acquisition, was announced for retirement on November 1, 2017 and shut down on April 10, 2018, with Google citing low interest among travel partners. Only the contract-only enterprise product remains. As of August 29, 2026 there is still no public replacement, so a scraper that reads the Google Flights results page is the practical route for independent developers and agents.
Is scraping Google Flights legal or allowed?
The data is public airfare information that any visitor sees without logging in, and none of the actors here require a Google account or bypass access controls. Google's terms of service discourage automated access, so the practical risk is Google throttling or blocking requests, not your credentials. Use the data for research, monitoring and analysis, keep request volumes reasonable, and consult a lawyer if you plan to redistribute fares commercially in your jurisdiction.
Why do the prices per 1,000 differ so much between these scrapers?
Because they bill different units. Memo23, automation-lab, skootle and Datapika charge per row, and all but Datapika add a fee per run start while Datapika adds Apify platform usage instead. Johnvc charges $0.03 per results page plus $0.01 per run, scrape.badger $0.012 per search regardless of rows, kaix $0.06 per 1,000 plus compute, and ScrapeBase $19.99 a month. Convert each to your actual rows per query before comparing; the worked example above does that.
Which Google Flights scraper has the best success rate?
On Apify's 30-day metric as of August 29, 2026, automation-lab, johnvc, makework36, leorochasantos and Scrape Sage show 100%, memo23 99.9%, scrape.badger 98.2%, Datapika 93.5%, kaix 92.8%, skootle 59.3% and ScrapeBase 0%. Weight that by volume: johnvc's figure covers roughly 29,800 runs in 30 days, leorochasantos 30 and Scrape Sage 51. A failed Datapika run pushes no rows and bills no itinerary events, and each search retries three times with backoff before giving up.
Can I get Google Flights data through MCP for an AI agent?
Yes. Every Apify actor can be exposed as a tool through Apify's MCP server, and both scrape.badger and Datapika document that path. For Datapika the tool URL is https://mcp.apify.com/?tools=fetch-actor-details,openclawai/google-flights-scraper. The agent sends airports, date, trip type and cabin and receives typed rows with price, airline IATA codes, segments, CO2 in grams and the echoed query inputs, so fan-out across dates can be merged without parsing HTML.
How many itineraries does a single Google Flights search return?
Google usually renders 30 to 80 itineraries for one route and date. Datapika returns up to your max_results, default 25 and maximum 200, and bills itinerary events only for rows delivered, so a typical search runs between 4.5 and 12 cents before platform usage. Automation-lab also caps at 200 per search. Per-search actors such as scrape.badger return the full page for one $0.012 charge, which is why that model wins on cost only when you want every result.