The best flight data APIs in 2026, compared
Published August 26, 2026 · Updated October 7, 2026
Short answer: FlightPowers' Google Flights Live API (mine) returns live Google Flights fares with Google's own low / typical / high price band and a round trip priced as one request. It is free to start: 10 searches a month on the RapidAPI BASIC plan, or 50 searches a day on the free MCP server with ads (free-trial.flightpowers.com/mcp, sign in with Google). Past that, PRO is $10 for 2,500 searches, about a sixth of SerpApi's price per search (their cheapest plan is $25 for 1,000). Best for price tracking, date scans and AI agents; it does not book. Air Scraper (cheaper per request, Skyscanner data), SerpApi (a Google scraping platform) and Duffel (a booking API) are the other three people actually shortlist in 2026, each quoted from its own live page below, including the one that beats mine on raw requests per dollar.
If you search for flight data APIs you get a wall of listings with similar names, similar screenshots and wildly different data underneath. This page compares the ones people actually shortlist.
I'll be honest: I'm starting with my own API. I run FlightPowers, so read the first entry as a vendor describing his product. Everything I say about the others is quoted from their own live pages, re-read on 2026-09-07, with the caveats stated inline, including the finding that is not flattering to me: on raw requests per dollar, one competitor is cheaper than I am at every tier. Nothing any of them charges changed between 2026-08-24 and this re-read; the one thing that did change is a SerpApi parameter, noted in its section.
1. Google Flights Live API (FlightPowers): mine
Source: Google Flights. Best for: price intelligence (alerts, calendars, "is this a good deal" features).
What I built it around: every result carries Google's own historical price band for
the route and dates (price_insights_low / price_insights_high) and Google's
verdict on the current fare: low, typical, or high. That is the difference
between an API that tells you a price and one that tells you whether the price is
good. A fare-alert product can trigger on that field on day one, with no price
history of its own. Beyond that: round-trip is a real paired-leg endpoint (one
request returns both legs with a combined total, not two one-way searches stapled
together), every result ships a working buy_link into Google Flights, per-minute
rate limits are published per plan so parallel date scans are a documented
capability, and every response carries an X-Search-Status header so "no flights"
and "the search failed" are different answers.
What it costs: free to start, two ways. The RapidAPI BASIC plan is 10 searches a month with no card, and the free MCP server with ads gives Claude, ChatGPT or Cursor 50 searches a day (250 a month) after a Google sign-in, no key at all. Then PRO is $10 for 2,500 searches, ULTRA $25 for 10,000 and MEGA $50 for 25,000.
Being fair about the cons, because everyone else on this page gets their pros: no
multi-city or open-jaw itineraries (one-way and round-trip only), no car rentals, and
the top public tier caps at 25,000 requests a month. All four cabins are covered:
seat_type 1 economy, 2 premium economy, 3 business, 4 first. Current plan prices live on /pricing; the
companion Booking.com hotels API is a separate listing on the same
account.
2. Air Scraper (sky-scrapper): the cheaper one
Source: Skyscanner. Best for: raw request volume, and scope beyond flights.
The headline finding, stated up front because it is true: Air Scraper is cheaper per request than my API. Prices read off its live RapidAPI listing on 2026-09-07 (identical to the 2026-08-26 read): free tier 20 requests/month with a hard cap, then $9.99/month for 10,600 requests, $49 for 55,000, $120 for 140,000, with published per-request overage of $0.002, $0.0015 and $0.001 respectively. That works out to roughly $0.00086–$0.00094 per request at every paid tier, against $0.0020–$0.0040 per request on mine. Its own listing describes it as an API that "helps to query prices for real-time flights, hotels, rental cars, etc… to create a travel website like the skyscanner.com or booking.com."
Two things make a "request" mean different amounts of work, and they are visible in
its own documented route list. First, Skyscanner addresses places by internal entity
IDs, not IATA codes, so its routes include searchAirport and getNearByAirports:
a user typing "New York" generally costs a resolution call before the search call.
(Mine takes IATA codes directly.) Second, its route list includes
searchIncomplete, searchFlightsComplete and searchFlightsWebComplete alongside
searchFlights, naming characteristic of Skyscanner's asynchronous model, where
you begin a search and poll for the finished set, and each poll is a request. I have
not measured how many calls a typical search resolves in, so I will not put a
multiplier on it. But one search is not reliably one request, so measure it on your
own workload before choosing on the per-request price. I also searched its listing
documentation for price-insight fields equivalent to Google's band and verdict and
did not find any (a statement about what the listing documents, not proof no such
field exists).
Genuinely fair pros: cheaper per request than mine at every published tier (that is a comparison with my listing, not a ranking of the Hub, which I have not priced end to end); flights, hotels and rental cars on one key; a price-calendar endpoint and an "everywhere" search as first-class endpoints (on mine you build those by scanning in parallel); multi-stop search, which I do not have at all; and a 140,000-request top tier against my 25,000. Against: rate limits are not published on its listing (its plan payloads carry no rate-limit value at all, checked 2026-09-07), and the data is Skyscanner's, which is not interchangeable with Google Flights coverage.
Consider Air Scraper if you need cars, multi-city, or volume past 25,000 requests a month. That list decides it, and this page should not talk you out of it.
3. SerpApi: the platform
Source: Google (many surfaces). Best for: teams that need more of Google than flights.
SerpApi is a general-purpose Google scraping platform where Google Flights is one of
several dozen engines on one subscription and one credit pool. From its own pages
(re-read 2026-09-07): plans run $25/month for 1,000 searches up to $275/month for
30,000 (roughly $0.0092–$0.025 per search), and round-trip shopping is a two-request
flow (their docs: to get returning flights "you need to make another request using a
departure_token"). Credits are fungible across all engines, and only successful
searches count toward quota, a genuinely customer-friendly billing policy.
One update since I first wrote this. SerpApi now documents a
selected_flights_json parameter that pins an itinerary segment by segment and takes
both an outbound and a return array, so repricing a round trip you already know
the flight numbers for is one request, not two. It does not change the shopping
case, which is what the two-credit arithmetic on /compare/serpapi
is about, but it is a real capability that was not there in August and it belongs in
their column.
Where it beats me, plainly: its price_insights object includes a price_history
array where we return a band; it returns per-itinerary carbon emissions; it supports
multi-city; and it advertises a legal indemnity of up to $2
million for search-engine scraping. The full side-by-side, with everything quoted
and dated, is at /compare/serpapi.
Consider SerpApi if flights are one feature of a product that also needs Google Search, Maps or Shopping data; if you want price history without accumulating it; or if your legal team wants the indemnity.
4. Duffel: the one that can actually sell a ticket
Source: airline inventory (NDC and their own IATAs). Best for: products where the purchase happens inside your product.
Duffel is not a data API and comparing it as one is a category error. It is a booking platform: order creation, ancillaries, and access to their IATA accreditation so you do not need your own. Its pricing (from duffel.com/pricing, re-read 2026-09-07) charges $3.00 per confirmed order, $2.00 per paid ancillary and lets search ride along free at a 1500:1 search-to-book ratio, with excess searches at $0.005. For a seller that converts, that model is excellent. For a non-booking workload (no orders means no free searches), every search costs $0.005, and its sign-up flow assumes you are selling flights. The full comparison is at /compare/duffel.
Consider Duffel if you are an OTA, corporate booking tool, agency or marketplace and the transaction happens inside your product. None of the data APIs on this page, mine included, can issue a ticket.
What is not on this list, and why
Several other listings turn up in the same RapidAPI search: flights-sky,
google-flights2, booking-com15, sky-scanner3 among them. Their pages did not
expose pricing in a form I could read reliably on 2026-08-24 or on the 2026-09-07
re-read, and I am not going to publish a price I did not verify.
Amadeus Self-Service used to be on every shortlist and is no longer a live
option. Checked on 2026-08-24 and re-run on 2026-09-07 with identical results: its
portal and pricing pages 301-redirect to the Amadeus homepage, the test sandbox host no longer resolves, all 20 amadeus4dev GitHub
repositories are archived, and Amadeus's own developer-guides repository states:
"The Amadeus for Developers Self-Service offer has been deprecated." Only the
Enterprise portal remains, which generally requires IATA/ARC accreditation. If you
are coming from there, the migration mapping is at
/compare/amadeus.
What are the best APIs for tracking flight status and prices?
Flight status (delays, positions, schedules) and flight prices are two different products: for prices, the four APIs on this page; for status, a reference feed such as aviationstack or FlightAware AeroAPI, both listed with their quoted plans in the travel data API category map; FlightPowers does prices, not status.
How to choose in five minutes
The free tiers make comparison pages, including this one, the wrong way to decide:
- Subscribe to two free tiers.
- Run your three hardest routes: the small regional airport, the multi-stop long-haul, the one your users complain about.
- Count requests per completed search, not requests.
- Check whether the response has what your feature needs. If your product must say "this is a good price," check for a price band.
- Then compare cost per completed, usable search.
curl -X POST https://api.flightpowers.com/v1/flights/oneway \
-H "x-api-key: $FLIGHTPOWERS_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"from_airport": "JFK",
"to_airport": "LHR",
"departure_date": "2026-11-12"
}'
The full walkthrough of mine is How to get real-time Google Flights data; a key comes from the RapidAPI listing, free tier included (10 requests/month, hard cap: verifies your key, doesn't evaluate).
Related
- Flight API benchmark 2026: four Google Flights APIs on the same three searches, timed, with raw numbers
- FlightPowers vs SerpApi: the full quoted side-by-side
- FlightPowers vs Duffel: data API vs booking platform
- Amadeus Self-Service vs FlightPowers: the migration path
- Price Insights API: the field this whole ranking hinges on, proven with a captured run
Decide on your routes, not my copy
Run your three hardest routes and check the response for a price band and a verdict. Free tier on RapidAPI, no card to try.
Free tier: 10 requests/month. No card to try.
Docs: the Flights API and the MCP servers.