Airbnb Scraper — Prices, Availability Calendar & Listings avatar

Airbnb Scraper — Prices, Availability Calendar & Listings

Pricing

from $4.00 / 1,000 listing rows

Go to Apify Store
Airbnb Scraper — Prices, Availability Calendar & Listings

Airbnb Scraper — Prices, Availability Calendar & Listings

Airbnb listings with prices for your dates and guests, by place, URL or listing id. One row per listing: nightly rate, stay total, rating, bedrooms, coordinates, Superhost, photos. Optional: 24-month availability calendar with occupancy, reviews. Free: amenities, house rules, host, price breakdown.

Pricing

from $4.00 / 1,000 listing rows

Rating

0.0

(0)

Developer

Tedj MEABIOU

Tedj MEABIOU

Maintained by Community

Actor stats

0

Bookmarked

101

Total users

63

Monthly active users

4 hours ago

Last modified

Features

Limited permissions

Share

An Airbnb scraper for Airbnb prices and availability: search any place or paste Airbnb URLs and get every listing's nightly rate and total for your dates and guests in any currency, its rating, location and badges — then, per listing, up to 24 months of availability calendar with occupancy data, the full listing details with the exact price breakdown, and reviews. Scrape Airbnb data as rows you own — the short term rental data behind market analysis, an Airbnb price tracker, or a rate check on your own listings — without an Airbnb API key, a login or cookies: it reads the same public endpoints the airbnb.com front-end uses. You pay per row you receive; a place Airbnb does not recognise, an empty search and every status row cost nothing.

Airbnb 房源、价格与可订日历:中文调用与语言设置

按地点、房源 URL 或 ID 获取房价与房源资料。calendarMonths 会增加日历输出;只查询入住报价时设为 0,maxReviewsPerListing 设为 0。

locale 控制来源提供的文本语言,currency 控制币种。示例 locale="fr"。不同市场、日期和人数必须保持一致后再比较;中文说明不表示所有房源原文都是中文。

调用需求参数与结果约定
小规模验证使用下面 JSON;先检查结果和状态,再提高上限
语言与市场按上面的真实字段选择,不要把中文说明当成自动翻译承诺
费用控制maxTotalChargeUsd 是 Apify 运行选项,不是输入 JSON 字段;按定价页的计费事件设置上限
数据完整性检查状态行和 SUMMARY;空结果、来源错误与完整数据集是不同情况
{
"locationQueries": [
"Lisbon, Portugal"
],
"checkIn": "30 days",
"checkOut": "33 days",
"adults": 2,
"currency": "EUR",
"maxListingsPerQuery": 1,
"calendarMonths": 0,
"maxReviewsPerListing": 0,
"locale": "fr",
"includeDetails": false,
"sessions": 1
}

Precios Airbnb / prix Airbnb / Airbnb Preise:价格按日期、入住人数和币种查询。

This localized quick start uses the existing input contract. API users should send the JSON as actor input and set maxTotalChargeUsd separately in Apify run options. Source coverage and translation depend on the platform, not the language of this guide.

What does this Airbnb scraper do?

A listing has a price only for a stay — dates and guests — and many scrapers quietly return a default figure, or listings from anywhere in the world when the place name is not understood. This one prices exactly the stay you ask for and tells you when it could not:

  • listing rows — for a place search, a pasted search URL or listing ids: total for the stay, base nightly rate and effective per-night price, discounts, rating and review count, bedrooms/beds/baths, coordinates, Guest favorite / Superhost badges, free cancellation, pictures. Airbnb answers about 280 listings per search; ask for more (maxListingsPerQuery up to 10,000) and the run covers the whole area in slices it answers in full — every listing of a city, each once.
  • calendar rows — one per listing: day-by-day availability for the next 1–24 months (available, available_for_checkin, min_nights…) plus an occupancy summary for the next 30/90/365 nights, the next available date and the typical minimum stay.
  • Listing details on request: description, amenities, house rules, the cancellation policy for your dates with its refund steps, check-in and checkout times, category ratings, host, capacity, sleeping arrangement, and the exact price breakdown for the dates — including "not available for these dates" when that is the answer.
  • review rows — newest, highest, lowest or most relevant first, with reviewer, date, text and host reply.
  • status rows — one per place, URL or listing: what Airbnb searched (resolved_place), how many rows were priced and charged, unknown_area, no_results or an error. Never charged.

Airbnb scraper for AI agents: which input answers which request?

Use it when the user wants Airbnb listings with the price for given dates and guests — a city or area, a pasted search URL, or specific listing ids — once or on a schedule; and, per listing, the availability calendar with occupancy, the details (amenities, house rules, host) or a few reviews. Do not use it for occupancy by month across a market (kestrel/airbnb-occupancy-scraper bills one small row per listing and month), for every review of a listing or a rating filter (kestrel/airbnb-reviews-scraper), for host contact details (not collected), or for hotels on booking sites (kestrel/google-hotels-prices).

The user asks forInput
Prices in a city for dates{"locationQueries": ["Lisbon, Portugal"], "checkIn": "2027-03-12", "checkOut": "2027-03-15", "adults": 2}
Every listing in a city, the whole marketadd "maxListingsPerQuery": 10000 — above 280 the run splits the search into slices Airbnb answers in full (Lisbon: about 6,100 for one stay); SUMMARY coverage shows Airbnb's own count next to what was delivered
Entire homes under 200 a night with 2 bedroomsadd "roomTypes": ["entire_home"], "priceMax": 200, "minBedrooms": 2, "currency": "EUR"
Amenities, house rules, host, cancellation policy, check-in and checkout times, price breakdownadd "includeDetails": true (no extra charge)
Availability calendar, occupancy, when is it freeadd "calendarMonths": 12 (one event per month per listing)
Prices and the latest reviews togetheradd "maxReviewsPerListing": 20, "reviewsSort": "most_recent"
These exact listings (portfolio, comp set){"listingIds": ["26119973", "1663639623527062900"], "checkIn": "30 days", "checkOut": "33 days"} — details come free
Every listing of a host or property manager{"hostIds": ["25841048"], "maxListingsPerQuery": 1000, "checkIn": "30 days", "checkOut": "33 days"} (or the profile URL https://www.airbnb.com/users/show/25841048 in startUrls) — one listing row each, priced for the stay
The listings from a search the user built on the site{"startUrls": ["https://www.airbnb.com/s/Paris--France/homes?checkin=2027-03-12&checkout=2027-03-16&price_max=200"]}
The price of listing links the user pasted, for the dates in each link{"startUrls": ["https://www.airbnb.com/rooms/26119973?check_in=2027-03-12&check_out=2027-03-15&adults=3"]} — each link is priced for its own dates and guests; a link whose stay Airbnb does not price is delivered free
Track prices weekly without editing the input"checkIn": "30 days", "checkOut": "33 days" and an Apify schedule
Tell me when new listings appear (scheduled)add "onlyNew": true, "monitorKey": "lisbon-new" — each run delivers and bills only listings no earlier run of that monitor delivered; with "maxListingsPerQuery": 10000 it watches the whole area
Listing data without dates (default displayed price)leave checkIn and checkOut empty; every listing row is billed
Prices in the destination's currency, names in French"currency": "EUR", "locale": "fr"
A family with a baby and a dog"adults": 2, "children": 1, "infants": 1, "pets": 1
A spending limitthe Apify run option maxTotalChargeUsd; maxListingsPerQuery, calendarMonths and maxReviewsPerListing bound the rows
The 40 cheapest listings"sortBy": "price_low_to_high", "maxListingsPerQuery": 40 — Airbnb has no price sort, so the run reads up to 3× the listings (at most 280), sorts them by stay total and delivers and bills the 40 cheapest in order; priceMax (per night) narrows the search first
Superhost only, rating 4.8 or better"superhostOnly": true, "minRating": 4.8 — applied before billing; a filtered-out listing costs nothing and the run reads further pages to fill your count
Listings rated between 3 and 4.3, the middle of the market"minRating": 3, "maxRating": 4.3, "maxListingsPerQuery": 10000 — a rating range, applied before billing (unrated listings drop out), over the whole area: Airbnb ranks top-rated listings first, so the first pages of a search rarely hold any (Asheville: 105 read, none in range, run ViVab3dh7iWMYGTzN; the whole area gave 47 of 3,740 in 73 s, run Mb2eQRXoLctq1HNZl). Add "maxItems": 20 to stop at 20
The minimum stay of each listingadd "calendarMonths": 1 — every listing row then carries min_nights (Airbnb's minimum stay for its check-in, read from the check-in's own month when it lies past the months asked, free) and min_nights_typical; the calendar row has it for every day
With a pool, hot tub, wifi, free parking…"amenities": ["pool", "hot tub"] (names or Airbnb ids; a listing must have all of them)
Free cancellation, Instant Book or Guest favorite only"freeCancellation": true, "instantBook": true, "guestFavorite": true (Airbnb's own filters)
Apartments only, houses only, hotels"propertyTypes": ["apartment"] (house, apartment, guesthouse, hotel); entire place vs private room is roomTypes
A neighbourhood drawn on the map"swLat": 38.70, "swLng": -9.16, "neLat": 38.72, "neLng": -9.13 — alone, or with a place to narrow it
This coming weekend, every week, no edits"checkIn": "friday", "checkOut": "sunday" (or "weekend" / "weekend") on a weekly schedule — each run prices the next Friday-to-Sunday; "checkOut": "3 nights" counts from the check-in
A nightly price for every day of the next 90 days, a price calendar"nightlyPriceDays": 90 — a night_price row per bookable check-in date with the price of the shortest stay Airbnb allows from it (add-on, from 2026-12-02)
Who hosts these listings, how big their portfolio is"includeHostDetails": true — one host row per host: Superhost, verified, languages, years hosting, rating, review count, number of listings (add-on, from 2026-12-02)

Where the bill is decided: maxListingsPerQuery (default 40) listing events per place search; calendarMonths and maxReviewsPerListing default to 0, so nothing beyond listing rows is billed unless asked; includeDetails is free. A listing the site cannot price for the dates is delivered with price_total: null and is not billed; without dates every listing row is billed at its displayed price.

Rows come in six types. listing: id, url, name, title, property_type, area, price_total, price_nightly, price_nightly_effective, price_original, price_nights, price_breakdown, price_taxes, price_total_with_taxes, currency, currency_mismatch, rating, reviews_count, bedrooms, beds, bathrooms, lat, lng, is_superhost, is_guest_favorite, free_cancellation, pictures, with calendarMonths also min_nights and min_nights_typical, and with includeDetails (or from listingIds, listing URLs and hostIds) also description, amenities, house_rules, cancellation_policy, cancellation_policy_id, cancellation_milestones, check_in_time, check_out_time, check_in_out, rating_breakdown, host, person_capacity, sleeping, available, unavailable_message. calendar: months, days (one entry per day with available, available_for_checkin, available_for_checkout, min_nights, max_nights), occupancy_30, occupancy_90, occupancy_365, next_available, min_nights_typical. review: review_id, rating, text, text_localized, language, created_at, reviewer_name, reviewer_location, host_reply, stay_type. night_price (add-on): id, date (check-in), check_out, nights, price_total, price_nightly_effective, price_nightly, currency. host (add-on): host_id, host_name, host_url, is_superhost, identity_verified, location, languages, about, years_hosting, rating, reviews_count, listings_count, host_listings (a sample). status: one per place, URL or listing id with status (ok, unknown_area, no_results, not_public, rounded_id, error), resolved_place, listings, priced, calendars, reviews, error, what a search left out (left_out, left_out_reason, note) and with onlyNew, already_seen; never charged. An unknown place or an empty search is a status row that says why, not a failed run. Blocked and booked nights look the same in the calendar: occupancy_* counts both. is_superhost is confirmed true on every row it carries; false on a search row is a lower bound, because a card that shows "Guest favorite" hides the Superhost badge — includeDetails (free) fills the true value from the listing page, and superhostOnly confirms each host before billing. On hotel cards title is the hotel's name and property_type is "Hotel".

Airbnb prices for exact dates — nightly rate, total and discounts

Every listing row carries price_total (what Airbnb displays for the stay after discounts), price_original (before discounts), price_nightly (the base nightly rate from the site's own breakdown) and price_nightly_effective (total ÷ nights), in the currency you chose, with price_breakdown lines such as "3 nights x $171.73", "Early booking discount −$89.73" and "Taxes". Every price is on one basis, the stay total before taxes with every fee Airbnb shows a logged-out guest, as US visitors see it. Visitors from the EU and UK get the total with taxes instead (Lisbon's tourist tax: $26 on a $305 stay), and proxy pools mix countries, so a scraper that copies the displayed total gives some rows with taxes and some without for the same input; here a total with taxes is turned back into the total before taxes from its own lines (a UK and a US exit priced 29 of 29 Lisbon listings identically, 2026-11-18) and, when Airbnb itemized them, the taxes come beside it as price_taxes and price_total_with_taxes. A price Airbnb gives in another currency than currency (it answers a currency it does not offer, such as GEL, in USD) is never delivered as yours: that row has no amounts, names Airbnb's currency in currency_mismatch, and is not charged. Without dates you get each listing's default displayed price and price_nights as Airbnb chose it. Prices depend on guests exactly as on the site, so set adults, children, infants and pets. An Airbnb nightly rate estimate for a market is one search — an Airbnb price calculator without the guesswork: the median price_nightly_effective of the rows.

Airbnb availability calendar and Airbnb occupancy data

calendarMonths: 12 adds a calendar row per listing: 365 days with availability flags and minimum/maximum nights, and occupancy_30, occupancy_90, occupancy_365 (share of nights not available — blocked or booked, the calendar does not tell which), available_90, next_available, min_nights_typical. This is the Airbnb availability API people search for and the company does not offer: an Airbnb calendar API you call with listing ids. Run it on a schedule and save each run's dataset externally to build an availability and price history for a market or a portfolio: the short term rental data API investors and revenue managers otherwise buy as AirDNA-style subscriptions, at a few cents per listing. Airbnb occupancy rate by city, by neighbourhood or by bedroom count is a group-by on the calendar rows.

Does Airbnb have an API? Airbnb API vs Airbnb scraper vs AirDNA alternative

Airbnb's only API is a closed partner program for property-management systems (listings, bookings, messages) — there is no Airbnb API for prices, availability or occupancy, no API key to request, and "Airbnb API access" for analytics does not exist. What developers actually use is one of two things:

Airbnb data vendors (AirDNA, AirROI, Airbtics, Mashvisor, Rabbu)This Airbnb scraper
DataAggregated, modelled estimates by market; monthly or weekly refreshRaw rows per listing, dates and guests, fetched when you run it
CoverageTheir markets and their listing sampleAny place Airbnb serves, any listing id, any currency and locale
PriceSubscription per market or planCents per row; empty searches free
OccupancyModelled from calendars and reviewsAvailability per night from the listing's calendar, with the caveat that blocked nights look booked
HistoryTheirsYours: every run appends dated rows
AccessDashboard or paid APIApify API, scheduler, webhooks, n8n/Make/Zapier, MCP for AI agents

So as an AirDNA alternative it is not a dashboard: it is the Airbnb data API layer you build a dashboard, a model or a spreadsheet on, free of a per-market subscription. If you searched for "Airbnb scraper Python", "Airbnb scraper free" or "Airbnb scraper Apify": the Python and JavaScript calls below are complete, Apify's free plan covers small runs, and the maintenance — the site rotates its query hashes and formats — is done for you, with a daily canary run and status rows instead of silent zeros.

Airbnb price tracker and Airbnb price history

Schedule the actor with relative dates ("checkIn": "30 days") and every run prices a stay at the same lead time. Save each run's dataset externally to build an Airbnb price history per listing — price_total and price_nightly_effective by fetched_at — and, with calendarMonths, an availability history. Connect a run-finished webhook to n8n, Make or Zapier for an Airbnb price tracker that alerts on drops in a shortlist, on a competitor undercutting you, or on new availability. For a portfolio, put your listing ids in listingIds and your competitive set in a second run; each listing is delivered and billed at most once per run, even across overlapping stays or guest counts.

Airbnb market data: comps, revenue and occupancy for investors

To scrape Airbnb listings for a market — a city, a neighbourhood, a coastline — combine a place search with calendarMonths: the listing rows give the price distribution for the dates (use price_nightly_effective for like-for-like nightly figures), the calendar rows give occupancy for the next 30, 90 and 365 nights per listing, and rating, reviews_count, bedrooms, room_type and coordinates let you segment. Airbnb comps for an underwriting: search the neighbourhood with the same bedrooms and room type, take the median price_nightly_effective and occupancy_90. A displayed-price × unavailable-nights calculation is a scenario, not observed revenue: blocked nights and sampled stay prices cannot establish bookings or realized earnings. Airbnb answers about 280 listings per search: set maxListingsPerQuery above 280 (up to 10,000) and the run covers the market itself (see "Every Airbnb listing in a city"); the status rows give Airbnb's own count (result_count), the pages read and the place resolved.

Airbnb reviews scraper

maxReviewsPerListing adds review rows per listing (a reviews scraper without a separate run) — newest first by default, or highest, lowest or most relevant — with the rating, text (and the site's translation when it exists), date, reviewer first name and location, the host's reply and the stay-type tag. reviews_total carries the listing's total review count so you know how much you sampled. To scrape Airbnb reviews for one property, give its id in listingIds with maxReviewsPerListing: 200.

Every listing of an Airbnb host

hostIds (numeric ids, or profile URLs such as https://www.airbnb.com/users/show/25841048, also accepted in startUrls) reads the host's public listing list — the one behind "View all listings" on their profile, 100 per request, so a manager with 590 listings takes 6 requests — and prices each listing for your stay like listingIds, with the free details: one listing row per listing (host_id set), billed as a listing row, up to maxListingsPerQuery per host (default 40; up to 10,000). One status row per host gives result_count. Use it to price a property manager's whole portfolio, or a competitor's, on a schedule. With includeHostDetails the host row also carries that whole list (host_listings) at no extra charge.

Every Airbnb listing in a city: whole-area coverage

Airbnb shows about 280 listings per search, while a city holds thousands (Lisbon 6,133 for a December stay, Porto 4,888). Set maxListingsPerQuery above 280 — up to 10,000 per place or search URL — and the run covers the whole area itself: it reads Airbnb's own count on the first page, splits the search into price bands along Airbnb's price slider, halves any band that still counts more than 270, then splits by room type, and a map area into quadrants, until every slice is one Airbnb answers in full; it reads the slices in parallel, de-duplicates them (Airbnb's price filter is fuzzy by about 2% at a band edge, and it re-ranks every request) and delivers each listing once, in price-band order. Rows bill on the usual listing event, maxItems and your spending limit cap the run exactly, and calendars and reviews follow every delivered listing. With sortBy the bands are read cheapest (or dearest) first and each one sorted. The status row and SUMMARY coverage report Airbnb's count (result_count), the listings delivered, the slices read and any listings Airbnb would not page to (truncated: a place search with thousands of listings at one price and room type). Timing: about 40 search pages per 1,000 listings, a few minutes for a whole city with the default 4 sessions.

{ "locationQueries": ["Lisbon, Portugal"], "checkIn": "30 days", "checkOut": "33 days", "maxListingsPerQuery": 10000, "currency": "EUR" }

maxItems on its own still lifts the per-search default to 280 only; whole-area coverage is always asked for with maxListingsPerQuery.

Airbnb nightly prices for every date (add-on, from 2026-12-02)

Airbnb's calendar no longer shows a price per day, so nightlyPriceDays: N rebuilds it: for each delivered listing it reads the calendar, takes every bookable check-in date in the N days from checkIn (from today without dates) and asks Airbnb for the price of the shortest stay it allows from that date — its minimum nights, so a listing with a 3-night minimum is priced as 3-night stays. Each answer is a night_price row: date, check_out, nights, price_total (with Airbnb's fees, for your guests and currency) and price_nightly_effective (total ÷ nights), billed per priced night ($0.0015, Store discount tiers apply). Booked or blocked dates and dates Airbnb will not quote cost nothing; the calendar is shared with calendarMonths when you ask for both. Use it for a price calendar, seasonality, the price of New Year's Eve against a normal week, or a revenue model from occupancy_* × nightly prices. The price takes effect on 2026-12-02 (Apify's notice period for a new event); until then a run that asks for it skips it, charges nothing for it and says so in its status.

{ "listingIds": ["1679053245445309369"], "nightlyPriceDays": 90, "calendarMonths": 3, "currency": "EUR" }

Airbnb host details (add-on, from 2026-12-02)

includeHostDetails: true adds one host row per host of the delivered listings, read from the host's public Airbnb profile: name, Superhost, identity verified (and since when), where they live, languages, about text, years hosting, rating and review count, the number of listings they manage and a sample of them with ratings — the professional-manager check (one Lisbon host manages 590 listings) that comps and market studies need. Each host is fetched and billed once per run, whichever listing led to it; it also fills the listing details (free), because the host's id comes from the listing page. No contact details exist in the public profile, so none are returned. Billed per host row ($0.002, Store discount tiers apply); the price takes effect on 2026-12-02, and until then the run skips it, charges nothing for it and says so.

Free on every listing row with details

Everything below comes from the listing page the run already reads for includeDetails, listingIds and hostIds (and its "Things to know", one more small read), so it costs nothing extra: the cancellation policy for your dates (cancellation_policy, e.g. "Free cancellation before January 10. Cancel before check-in on January 15 for a partial refund.", with cancellation_milestones: each refund step, when it applies and what it returns, and Airbnb's policy id), check-in and checkout times (check_in_time "4:00 PM", check_out_time "11:00 AM", and check_in_out with the lines as Airbnb shows them, such as "Self check-in with lockbox"), the registration or licence number when the host publishes one (registration_number, e.g. "74813/AL" in Lisbon), safety and property disclosures (safety: security cameras, carbon monoxide and smoke alarms, noise, parking), whether children and infants are allowed, co-hosts, the host's response rate and response time, languages and Airbnb-managed flag (inside host), the full photo gallery (pictures, often 30+), amenities, house rules (capacity, pets, smoking, events, quiet hours), sleeping arrangement, category ratings and the exact price breakdown. Airbnb states a cancellation policy only for a stay, so without dates cancellation_policy stays empty; the check-in times come either way.

Monitor new Airbnb listings

Set onlyNew: true and a monitorKey on a scheduled run and each run delivers and bills only the listings that earlier runs of that monitor did not deliver: the first run returns the current listings and remembers them, later runs return what is newly listed, or newly matches your filters. A listing id, or a host's listing, delivered before is not fetched again. The memory is a key-value store named airbnb-prices-calendar-seen in your Apify account (record seen-<monitorKey>; without a key, one named after the input's places, URLs, listing ids and hosts), and only delivered listings join it, so one a spending limit kept out comes back next run. To catch every new listing of a city, run it in whole-area mode (maxListingsPerQuery above 280) or on a map area: a 40-listing search sees only the listings Airbnb ranks first, and Airbnb re-ranks every request. These are listings newly seen by the monitor, not confirmed new openings; status rows count the skipped ones in already_seen, SUMMARY in only_new.

{ "locationQueries": ["Lisbon, Portugal"], "checkIn": "30 days", "checkOut": "33 days", "maxListingsPerQuery": 10000, "onlyNew": true, "monitorKey": "lisbon-new-listings" }

Airbnb filters: amenities, property type, free cancellation, Instant Book, Guest favorite and map area

The filters on Airbnb's own filter panel are inputs here, applied by Airbnb before the cap so every returned listing matches and nothing filtered out is billed. amenities takes names or Airbnb's ids and keeps listings that have all of them (["pool", "hot tub"]: 18 of Lisbon's 6,133 listings for one December stay; names: wifi, kitchen, air conditioning, heating, washer, dryer, pool, hot tub, free parking, ev charger, dedicated workspace, tv, self check-in, crib, king bed, bbq grill, breakfast, indoor fireplace, smoking allowed, step-free access and the other accessibility features). propertyTypes is Airbnb's property-type filter (house = homes, cabins and townhouses; apartment = apartments, condos and lofts; guesthouse = guest suites; hotel = hotels and B&Bs), a different axis from roomTypes (entire place, private room, shared room, hotel room). freeCancellation (Lisbon: 6,133 → 4,471), instantBook (→ 5,705) and guestFavorite (→ 1,843) are Airbnb's booking-option and top-tier filters. A map area (swLat, swLng, neLat, neLng) searches exactly that box, as when you move the map on airbnb.com: alone it is a search of its own, with a place it narrows the place. Airbnb labels the cards of a box searched alone with no type line, so their title, property_type and area stay empty: add the place for those, or includeDetails for room_type. These filters apply to place searches and override a pasted search URL's own value; price, roomTypes and room counts stay place-search filters, and minRating, superhostOnly and sortBy apply to every listing. A pasted URL keeps every filter of the site's panel, including ones without an input field here (Luxe, host language, Beachfront, neighbourhood).

Search URL context and overlapping searches

A search URL keeps its embedded stay and guest filters. Explicit checkIn and checkOut inputs override its dates. URL guest values remain authoritative for that search; when absent, its existing URL-parser defaults apply (2 adults, no children, infants or pets). Listing and status dates and optional detail-price requests now use that same effective search context.

A listing URL keeps its own stay too: https://www.airbnb.com/rooms/26119973?check_in=2027-01-15&check_out=2027-01-18&adults=3 is priced for those three nights and three adults, and its row and status row carry them. Explicit checkIn and checkOut override its dates, as for search URLs; its guests win over adults and the other guest inputs, which fill in what the link does not say. A link with dates is billed only when Airbnb prices that stay; dates that are past or not a stay are dropped with a note on the status row, and that row is free. (Version 1.14 priced such links undated, for 2 adults, and billed them: run WxhTeGAOafXPwteqK; test test_a_listing_url_is_priced_for_its_own_dates_and_guests.)

A listing is delivered and billed at most once per run, including its calendar/review add-ons, even when overlapping searches request different stays or parties. The first processed occurrence wins; concurrent searches need not finish in input order. Use separate runs to compare the same listing across distinct stays or guest counts, and inspect the emitted stay/guest fields.

Join a stay quote to an availability calendar

This call asks for five Lisbon listings for one three-night stay, with amenities and two calendar months of availability. It costs at most $0.045 for five priced listings ($0.025) plus ten calendar-month events ($0.020), before the $0.05 run cap. It asks for no reviews.

{
"actor": "kestrel/airbnb-prices-calendar",
"input": {
"locationQueries": ["Lisbon, Portugal"],
"checkIn": "2027-03-12",
"checkOut": "2027-03-15",
"adults": 2,
"currency": "EUR",
"maxListingsPerQuery": 5,
"includeDetails": true,
"calendarMonths": 2,
"maxReviewsPerListing": 0
},
"callOptions": {"maxTotalChargeUsd": 0.05, "timeout": 180},
"waitSecs": 30
}

This is a call-actor envelope; only input belongs in the Actor input editor. If it is still running, poll get-actor-run, then fetch the default dataset with get-dataset-items. Group rows by type. Join each calendar row to its listing row using id; a two-month calendar is one dataset row containing days[], but two billing events. Five complete results can therefore be five listing rows, five calendar rows and one free status row. A missing calendar is not zero availability: check calendar_errors, returned months, and SUMMARY.limit_hit.

listing.price_total and price_nightly_effective describe the quoted stay. Calendar months start in the current month, independently of checkIn; daily calendar entries supply availability/stay rules, not nightly prices (for a price per date add nightlyPriceDays). Do not spread the stay price across all calendar dates or equate unavailable days with confirmed bookings. occupancy_* uses the returned days in its named window; check nights_30, nights_90 or nights_365 before comparing differently sized calendar samples.

To monitor this same March stay, keep its absolute dates and save each run's dataset with fetched_at. Relative dates such as 30 days / 33 days produce a moving three-night stay, not the same check-in dates. Availability-based revenue calculations are scenarios, not observed earnings.

Where other scrapers fail, and what this actor does instead

What users of the most used Airbnb scrapers on the Store report in their issues and reviews (read 2026-11-18), and what happens here instead. Each line names the test in this actor's suite or the platform run that shows it.

Users reportThis actor
Prices only in GBP, or "random currencies per run", whatever currency was asked (two 2★ reviews, open issues)Every price is checked against currency. One in another currency is asked again from other IPs; if Airbnb still will not price in it (it answers GEL in USD) the row arrives without amounts, names Airbnb's currency in currency_mismatch and is free. Test test_a_currency_airbnb_does_not_offer_delivers_listings_free_and_says_so; run WVgnfhkJIwMOfiMg9 (10 listings asked in GEL: delivered free, explained on the run card), where run WiW525jTSlsZ6vbJr of the previous version delivered USD prices labelled GEL.
"Sometimes includes taxes and sometimes doesn't for the same input"One basis, the total before taxes, whatever country the proxy exit is in; the taxes beside it when Airbnb showed them. Runs YT44WmxEfJfWiEYAF (default proxy), vUa72On2eXwgXrSm5 (UK) and JydofS8r0f19zpZp6 (Germany) of the previous version priced every shared listing apart by Lisbon's tourist tax; now runs 7AXGVXmSHXnPk7daa (default proxy) and Rk0HoxQfIaFXcdYJo (UK, every total shown with taxes) price all 30 shared listings identically, and run 8uA4fHcMGmHHIbbmX reads French totals with taxes the same way (test test_a_total_with_taxes_comes_back_on_the_basis_us_visitors_get).
Empty answers to the same call that worked a minute agoAn empty first page that Airbnb did not count as empty (a search that matches nothing answers a count of 0) is asked again from other IPs before a search is reported empty. Test test_an_empty_first_page_is_asked_again_before_no_results_in_every_search_mode.
"Only 41 results returned when Airbnb shows 478 listings"Whole-area mode delivers every listing Airbnb counts, each once: Prague 2,288 for Airbnb's own count of 2,224 (run yO93wSY9tdQuWRhUr). Test test_whole_area_delivers_every_listing_once_and_charges_exactly_them.
"No results — Proxy error 597 UPSTREAM407"3 refusals in a row move the run to residential proxies of the same country: run uvJMPT2GU5nPDwEW7, with a proxy country that has no datacenter IPs, succeeded. Test test_a_proxy_country_without_datacenter_ips_escalates_to_residential_ips_of_that_country.
Runs that time outNo new work starts close to the run timeout; the run ends green with everything it delivered. Test test_near_the_run_timeout_nothing_more_goes_out_and_the_run_ends_green.
Charged for listings from the wrong areaA place Airbnb does not recognise gets a free unknown_area status row. Test test_unknown_place_charges_nothing.
Runs "still show Succeeded" while every listing failedA query Airbnb retired fails the run loudly, after delivering and billing everything that did not need it. Test test_a_retired_query_fails_the_run_loudly_after_delivering_and_charging_the_rest_once.
"Add minimum nights to output" (open request)With calendarMonths, every listing row carries min_nights and min_nights_typical from the calendar it already reads: runs xOYh1G4r8PjDMkSwr (a search) and AupAd9nzm9PxN5sic (listing ids). Test test_calendar_months_put_min_nights_on_listing_rows_from_the_same_calendar.
"Target Median Rated scores 3-4.3" (open request)minRating with maxRating is a rating range, applied before billing: run aAOHDhwTIgPV4xMbb (10 listings rated 4.5–4.85). Airbnb ranks top-rated listings first, so a 3–4.3 range needs the whole area (maxListingsPerQuery: 10000): 47 of Asheville's 3,740 (run Mb2eQRXoLctq1HNZl). A range that finds none in the pages it read says so and names that setting (test test_a_rating_range_that_finds_none_in_its_pages_says_so_and_points_to_whole_area_mode).
"I wish it would tell me when new listings appear"onlyNew with a monitorKey: each scheduled run delivers and bills only listings no earlier run of the monitor delivered. Test test_only_new_delivers_and_bills_each_listing_once_across_runs_of_a_monitor.
"calendar[].price is always null"nightlyPriceDays (add-on, from 2026-12-02) prices each bookable check-in for exactly its stay. Test test_night_prices_bill_one_event_per_priced_night.
"The only thing missing is the check-in and check-out times"On every row with details: check_in_time, check_out_time and the lines as shown, read signed out from the listing's "Things to know" (2026-11-18: on 9 of 9 listings in six languages). Test test_listing_rows_with_details_carry_the_cancellation_policy_and_check_in_times.

Questions users ask about other scrapers

The most repeated complaints in the issues and reviews of the 18 most used Airbnb scrapers on the Store (74 issue threads and 41 reviews read on 2026-11-18), answered for this actor. Each answer names the run or the test that shows it; every run is of 2026-11-18, on version 1.15 or, for behaviour 1.15 keeps, on 1.14, and "before" marks a 1.14 run showing what 1.15 fixed.

"The run says Succeeded but returned only ids", "my site is down since the scraper broke". Airbnb's four queries answered on 2026-11-18 (run WxhTeGAOafXPwteqK). A rotated query hash is picked up by itself; a query Airbnb retires fails the run loudly after every row that did not need it was delivered and billed, instead of a green run of empty rows (test test_a_retired_query_fails_the_run_loudly_after_delivering_and_charging_the_rest_once); a missing section leaves its fields empty and never drops the row (test test_parsers_survive_empty_and_malformed_payloads).

"Description, host or coordinates are missing", "the cancellation policy is empty". Details carry description, amenities, host, coordinates, ratings and house rules (6 of 6 rows in run WxhTeGAOafXPwteqK). The cancellation policy moved out of the listing page's answer in November 2026 and was empty on every row there; it is back, for your dates and with its refund steps (run 1irI7b9pL1iueqzbI: the policy, its refund steps and the check-in and checkout times on 5 of 5 rows; before, run woDyX79xYCQWiobsG: the policy on 0 of 5; test test_listing_rows_with_details_carry_the_cancellation_policy_and_check_in_times).

"Only returns GBP", "random currencies per run". Prices come in the currency asked whatever the proxy's country: 40 of 40 rows in USD through a UK exit (run DvhRPxCVlu5viu5yD), 20 of 20 in CAD (run sN6bquQPHc7GVfxEb). A currency Airbnb does not offer arrives without amounts, free, and the run card says so (run lne7GPe8hd3UlwB2M).

"totalPrice includes taxes in one run and not the next". One basis, the total before taxes, with the taxes beside it when Airbnb showed them: a UK exit and a default exit priced the same 26 listings identically (runs DvhRPxCVlu5viu5yD and upz8e4zkzZvonW99W).

"It returned listings from a completely different region, and charged me". A place Airbnb does not recognise is a free unknown_area status row; small towns resolve to themselves, a median 1 km from their centre (run hS96CA1JfI3AdbCFr). A pasted room URL returns that exact listing (run WxhTeGAOafXPwteqK).

"checkIn and checkOut are ignored for my listing URLs". A listing URL is priced for its own check_in, check_out and guests (run TiNPrvKDUBKZSKBDu; before, run WxhTeGAOafXPwteqK priced it undated for 2 adults and billed it). A link whose stay Airbnb does not price is free, and so is one whose dates are past (run ovzHZz0QIgELlDrxK). Test test_a_listing_url_is_priced_for_its_own_dates_and_guests.

"Only 41 results when Airbnb shows 478". Every search reads Airbnb's own count, and the run card says what a limit left out: "40 of the 262 listings Airbnb counts for Moulins, France: maxListingsPerQuery 40 stopped there" (run HiwCKMa8grwKZDC6Z; before, run j6OTuzJ59Hmw2ygcF said only "40 priced listings"). With maxListingsPerQuery: 300 the same town gave 224 of Airbnb's 226 (run bqpn1E253U2Xntf3T); above 280 the run covers the whole area.

"I asked for listings rated 3 to 4.3 and got nothing". Airbnb ranks top-rated listings first, so the first pages hold none: the run now says how many it read and that maxListingsPerQuery: 10000 scans the whole area (run TNEk6pV9Ny6l6h0l3; before, run ViVab3dh7iWMYGTzN said "0 priced listings"); the whole area gave 47 of 3,740 (run Mb2eQRXoLctq1HNZl).

"Add minimum nights", "check-in and check-out times are missing". With calendarMonths, every dated row carries min_nights, even for a check-in past the months asked (run wZLojarK4cuKQyzwT: 5 of 5 rows for a January check-in with calendarMonths: 1; before, run WxhTeGAOafXPwteqK: 0 of 5). Rows with details carry check_in_time and check_out_time (run 1irI7b9pL1iueqzbI).

"Does it take 18- or 19-digit ids?". Yes, as strings, or inside a /rooms/ URL. An id sent as a JSON number above 2^53 is rounded by JSON before the run starts, so it may name another listing: it is not fetched, and its free status row says "send listing ids as strings" (run 2ovo4BWnavUbs7vTp; before, run StxQnl1bLlsFhSPSI called it "not shown publicly").

"Charged me despite the limits", "my small test run expanded into a huge one". A search stops at 40 listings unless asked (run j6OTuzJ59Hmw2ygcF); maxItems is exact; a $1.00 limit delivered and billed exactly 200 listings (run sCndeyfiUvqlirmTD); runs that deliver nothing bill nothing (runs ViVab3dh7iWMYGTzN, hS96CA1JfI3AdbCFr). Switching from another scraper keeps your limits: a includeReviews: false beside maxReviewsPerListing: 100 bills no reviews (run GCMzmS1o64XdZXWHk; before, run 4gEmkmSSQP4tRc1Ic billed 63 reviews), and cirkit's includeCalendar: false beside calendarMonths: 6 bills no calendar (run WdNDlhCYvmObKqOSQ; before, run w1AR9P3hpGp55w2hL billed 6 calendar months). For a scheduled or looped call, set maxTotalChargeUsd per run and onlyNew so a repeated run bills only what is new.

"Proxy responded 597 UPSTREAM407", "timeout of 30000ms exceeded". A proxy country without datacenter IPs moved to residential proxies after 3 refusals and succeeded (run zMPRnAQ6SJDTuuqpI); a run with a 60-second timeout stopped new work at 34 s and ended green with every row it delivered (run sCndeyfiUvqlirmTD).

"Can't export the results", "the CSV has hundreds of columns". Calendar rows nest one entry per day, so the full CSV of a calendar run is wide. For one row per day, download https://api.apify.com/v2/datasets/<datasetId>/items?format=csv&unwind=days&fields=type,id,name,days (run wZLojarK4cuKQyzwT: 150 day rows for its 5 one-month calendars, next to the listing and status rows; keep type to tell them apart).

Input

FieldMeaning
locationQueriesPlaces as you would type them on the site: "Lisbon, Portugal", "Algarve", "Brooklyn, NY". Unknown places produce an unknown_area status row and cost nothing.
startUrlsSearch URLs (dates, guests and every filter of the site's panel in the URL are honoured) or listing URLs; plain strings or {"url": …} objects.
listingIdsListing ids to price, calendar, detail or review directly.
checkIn, checkOutYYYY-MM-DD, relative to the run date ("30 days", "in 6 weeks", "tomorrow") or a weekday ("friday", "next saturday", "weekend" = the first Friday after the run day). checkOut also takes "3 nights" and weekdays counted from the check-in ("sunday", "weekend" = the first Sunday after it). Leave both empty for default prices.
adults, children, infants, petsGuests; prices depend on them.
currency, localeAny ISO currency; site locale for names, descriptions and reviews.
priceMin, priceMax, roomTypes, minBedrooms, minBeds, minBathroomsPlace-search filters (prices per night in currency, compared by Airbnb as the stay total with fees ÷ nights; decimals work; 0 = no limit).
amenities, propertyTypes, freeCancellation, instantBook, guestFavoriteAirbnb's own filters (see "Airbnb filters" above); they also override a pasted search URL's value.
swLat, swLng, neLat, neLngA map area: all four, or all 0 (the default) for none.
maxListingsPerQueryDistinct listings per place or search URL, up to 10,000. Up to 280 is one Airbnb search; above it the run covers the whole area in slices (see below). maxItems alone lifts the default to 280, not beyond.
includeDetailsFetch each listing's page (details + exact breakdown; one extra call per listing).
calendarMonths0–24 months of availability per listing.
maxReviewsPerListing, reviewsSortReviews per listing and their order.
minRating, maxRating, superhostOnlyFilters applied before billing: a minimum and a maximum Airbnb rating (a range with both; unrated new listings drop out) and Superhost hosts only (Airbnb's own Superhost filter, then each host checked on its listing page). A filtered-out listing costs nothing; the run reads further pages to fill your count.
onlyNew, monitorKeyScheduled monitoring: deliver and bill only listings earlier runs with the same monitorKey did not deliver (see "Monitor new Airbnb listings").
sortByrelevance (default, Airbnb's own order), price_low_to_high or price_high_to_low. Airbnb has no price sort, so the run reads up to 3× the listings you asked for (at most 280), sorts them by stay total and delivers the top ones.
proxyConfiguration, sessions, perIpApify datacenter proxies work; defaults are fine. When the configured proxy refuses 3 requests in a row (proxy errors, 403, 429; for example a proxy country where Apify's datacenter pool has no IPs, such as PT), the rest of the run goes through Apify residential proxies of that country (US without one), at no extra charge to you.

Example: Airbnb prices for a weekend in Lisbon

{ "locationQueries": ["Lisbon, Portugal"], "checkIn": "2027-03-12", "checkOut": "2027-03-15", "adults": 2, "currency": "EUR", "maxListingsPerQuery": 80 }

Example: occupancy and prices for a market, weekly, always 30 days out

{ "locationQueries": ["Algarve, Portugal"], "checkIn": "30 days", "checkOut": "33 days", "adults": 2, "currency": "EUR", "maxListingsPerQuery": 200, "calendarMonths": 6 }

Example: your own listings and competitors — details, 12-month calendar, latest reviews

{ "listingIds": ["26119973", "1663639623527062900"], "checkIn": "30 days", "checkOut": "33 days", "includeDetails": true, "calendarMonths": 12, "maxReviewsPerListing": 20, "reviewsSort": "most_recent" }

Example: a pasted search with its own filters

{ "startUrls": ["https://www.airbnb.com/s/Paris--France/homes?checkin=2027-03-12&checkout=2027-03-16&adults=2&price_max=200&room_types%5B%5D=Entire%20home%2Fapt&min_bedrooms=2"], "currency": "USD" }

Example: this coming weekend, entire apartments with a pool and free cancellation, every week

{ "locationQueries": ["Lisbon, Portugal"], "checkIn": "friday", "checkOut": "sunday", "roomTypes": ["entire_home"], "propertyTypes": ["apartment"], "amenities": ["pool"], "freeCancellation": true, "currency": "EUR", "maxListingsPerQuery": 40 }

On a weekly schedule each run prices the next Friday-to-Sunday.

Example: the 20 cheapest Superhost stays rated 4.8 or better

{ "locationQueries": ["Lisbon, Portugal"], "checkIn": "30 days", "checkOut": "33 days", "maxListingsPerQuery": 20, "minRating": 4.8, "superhostOnly": true, "sortBy": "price_low_to_high" }

Only the 20 delivered listings are charged; the extra pages read for the sort and the host checks cost nothing.

Inputs from other scrapers work too

Inputs written for the other Airbnb scrapers on the Store run here as they are: their field names are read as ours when our own field is absent, and SUMMARY lists them under aliased_fields (every other unknown field is skipped and listed under ignored_fields, never a failed run). A row limit, a filter or a switch that saves money is never dropped: every such field in the public input forms of the 17 most used other Airbnb scrapers is read (2026-11-18; tests test_rival_limits_and_money_saving_switches_are_read_never_ignored, test_kaixs_max_location_deviation_drops_far_listings_before_billing). Not read, because nothing here matches them: request and concurrency settings (automation-lab's maxRequestsPerCrawl, maxConcurrency, timeoutMs), kaix's coordinate search (latitude, longitude) and cirkit's calendar start month; they are listed in ignored_fields.

Their field (scraper)Read as
maxResults (tri_angle, openclawai), maxListings (automation-lab), count (curious_coder)maxItems: total listing rows of the run (0 = their "no limit": nothing is capped)
startUrls as {"url": …} objects (tri_angle, automation-lab)startUrls
urls (curious_coder), searchUrl / roomUrls (openclawai), listingUrl (agenscrape)startUrls (search and listing URLs)
limitPerUrl (curious_coder)maxListingsPerQuery
checkInDate, checkOutDate (curious_coder)the stay, for URLs that carry no dates of their own (their meaning: a default)
scrapeDetail (curious_coder), skipDetailPages: false (automation-lab), mode: "room_details" (openclawai)includeDetails: true (free here)
scrapeAvailability (curious_coder) / mode: "room_calendar" (openclawai) / startDate + endDate (agenscrape)calendarMonths: 3 / 12 / the months through endDate
enrichWithPricing (agenscrape)nightlyPriceDays: the days from startDate to endDate (at most 365)
enrichUserProfiles (tri_angle)includeHostDetails
hostId with mode: "host_listings" (openclawai), host profile URLs (curious_coder urls)hostIds: the host's whole portfolio as listing rows
maxListings: 0 (automation-lab: no limit)maxListingsPerQuery: 10000, the whole area
a saved search URL whose dates are past (curious_coder's own example is dated 2024)searched without dates, with a note on the status row
locale spellings such as es-419 or zh-TW-US (tri_angle)accepted as they are
scrapeReviews (curious_coder) / mode: "room_reviews" with maxResults (openclawai)maxReviewsPerListing: 10 / maxResults (reviews per room in that mode, 50 without it)
placeType (openclawai)roomTypes
language (openclawai)locale (en-US-style locales work as they are)
proxy (curious_coder)proxyConfiguration
neLat, neLng, swLat, swLng, amenities, freeCancellation (openclawai), calendarMonths, priceMin, priceMax, minBeds, minBedrooms, minBathrooms, adults … (tri_angle, automation-lab)the same names here
location (kaix, onidivo, integrative_tangent), listings (kaix: listing URLs or ids, read only without a location, as kaix does)locationQueries; startUrls / listingIds
minPrice, maxPrice (kaix, onidivo: per night)priceMin, priceMax
propertyType (kaix: entire_home, private_room …), reviewsSortBy (kaix: MOST_RECENT …)roomTypes, reviewsSort
maxSearchPages (kaix), max_pages (codepoetry): result pages of 18maxListingsPerQuery = 18 × pages, so a search never reads more than it would there
includeReviews (kaix, memo23), include_reviews (codepoetry), includeCalendar (cirkit)the switch wins over the count beside it: false = no reviews (no calendar) fetched or billed even when maxReviewsPerListing: 100 or calendarMonths: 6 comes with it; true without a count = 20 reviews (6 months)
monitoringMode (memo23)onlyNew: only listings no earlier run of this input delivered
fullCoverage (memo23)maxListingsPerQuery: 10000, the whole area
months (cirkit's availability scraper), maxReviews (onidivo: reviews per home)calendarMonths, maxReviewsPerListing
onlyAvailableListings (lergassy: skip fully booked listings)each listing's calendar is read before billing (free when calendarMonths is 0: six months, not delivered) and a listing without a free night is never delivered or billed
boundingBox (integrative_tangent: ne_lat, ne_lng, sw_lat, sw_lng)swLat, swLng, neLat, neLng
maxLocationDeviation (kaix: km from the searched place)a distance filter before billing, from the median point of Airbnb's first page: a listing farther away is never delivered or billed
strictLocationMatch (memo23), extraData (onidivo)always on here (an unrecognised place is a free unknown_area row); includeDetails
snake_case names (codepoetry: location_queries, max_listings, check_in, min_rating, price_max, calendar_months …), mode: "full" / "search" (codepoetry)the same fields here; includeDetails true / false

Not covered: Airbnb Experiences (mode: "experience_search") end as an input problem with nothing charged.

Output

listing row (search; with includeDetails the same row also carries description, amenities, house_rules, cancellation_policy, cancellation_milestones, check_in_time, check_out_time, rating_breakdown, host, person_capacity, available):

{ "type": "listing", "id": "1018680572065141154", "url": "https://www.airbnb.com/rooms/1018680572065141154", "name": "Lisboa Bica Nature Fast Net AC Heating", "title": "Apartment in Lisbon",
"query": "Lisbon, Portugal", "check_in": "2026-10-03", "check_out": "2026-10-06", "nights": 3, "adults": 2, "children": 0, "infants": 0, "pets": 0, "currency": "USD",
"price_total": 426.0, "price_total_display": "$426", "price_original": 516.0, "price_nightly": 171.73, "price_nightly_effective": 142.0, "price_nights": 3,
"price_breakdown": [{ "label": "3 nights x $171.73", "amount": 515.19 }, { "label": "Early booking discount", "amount": -89.73 }, { "label": "Price after discount", "amount": 425.46 }],
"rating": 4.88, "reviews_count": 156, "bedrooms": 1, "beds": 1, "bathrooms": null, "lat": 38.70975, "lng": -9.14708, "badges": ["Guest favorite"], "is_guest_favorite": true, "is_superhost": false, "free_cancellation": false,
"pictures": ["https://a0.muscache.com/im/pictures/…"], "details": false, "fetched_at": "2026-08-26T09:10:00+00:00" }

calendar row:

{ "type": "calendar", "id": "11562247", "name": "Charming studio in Baixa", "months": 12, "first_day": "2026-08-01", "last_day": "2027-07-31",
"nights_30": 30, "available_30": 6, "occupancy_30": 0.8, "nights_90": 90, "available_90": 47, "occupancy_90": 0.478, "nights_365": 340, "available_365": 137, "occupancy_365": 0.597,
"next_available": "2026-08-29", "min_nights_typical": 3, "days": [{ "date": "2026-08-29", "available": true, "available_for_checkin": true, "available_for_checkout": true, "min_nights": 3, "max_nights": 1125, "bookable": null, "price": null }], "fetched_at": "…" }

review row: { "type": "review", "id": "11562247", "review_id": "1740402488245156304", "rating": 5, "text": "…", "created_at": "2026-07-29T07:09:04Z", "reviewer_name": "Sarah Sofia", "reviewer_location": "…", "host_reply": null, "language": "en", "reviews_total": 441 }

status row (per place, URL or listing_id; result_count is Airbnb's own count for the search, and whole-area searches add slices, splits, truncated, slice_errors): { "type": "status", "query": "Lisbon, Portugal", "listing_id": null, "status": "ok", "resolved_place": "Lisbon, Portugal", "canonical_url": "/s/Lisbon--Portugal/homes", "result_count": 6133, "pages": 2, "listings": 80, "priced": 80, "duplicates": 0, "calendars": 80, "calendar_months": 480, "reviews": 0, "error": null }

The Console's Listings, Calendar, Reviews and Run status tabs show the same rows as tables; every run's dataset downloads as JSON, CSV or Excel — an Airbnb dataset CSV for exactly the market, dates and listings you chose, instead of a stale Kaggle file.

How fast is it?

Measured on 2026-11-17 on the platform with the default 4 sessions (time from the run's start, container start included): 20 London listings for a stay in about 3 s in total; the Input tab's example (40 Lisbon listings) in 5–7 s; 5 listings with a 12-month calendar each in 4–5 s; a 300-listing whole-area run of Lisbon in 15–55 s over 15 runs, most of it Airbnb's answer time through the proxy IPs a run happens to get (the same four IPs took 16 s and 36 s an hour apart); Prague's whole market (2,288 listings) in about 90 s. The first rows land after the first search page, not at the end. For comparison, a 20-listing London run of the most used Airbnb scraper on the Store took 12 s. Add-ons that read one page per listing or date (includeDetails, nightlyPriceDays, hostIds) run about 4 requests per second; details are two requests per listing, its page and its "Things to know": 5 Lisbon listings with details took 7–17 s in three of four runs on 2026-11-18 and 40 s in one on slow proxy IPs, where version 1.14 took 4.5–11.5 s at the same moments without the second request (paired runs h5rkXaOvaJ2qTk6I9 / KHG5vRisvCleBnw4f and Wehz9Zu5qMBClKp4R / xshhiS4SXZpzbCXuy); raise sessions (up to 16) for big batches.

How much does it cost?

Pay per row: a listing row that carries a price, each calendar_month (a 12-month calendar row = 12 events), each review row, and with the add-ons each night_price row and each host row. Listings that could not be priced, unknown places, empty searches and status rows are free, and each listing is delivered and billed at most once per run even across overlapping stays or guest counts. Worked examples: 80 Lisbon listings with prices for a weekend = 80 listing events; the same 80 with a 6-month calendar = 80 + 480 calendar-month events; your 10 listings with details, 12-month calendars and 20 reviews each = 10 + 120 + 200 events. Platform compute for those runs is a fraction of a cent. Exact per-event prices are on this page's pricing tab.

Airbnb scraper in Python, JavaScript, curl, n8n, Make or an AI agent

"Airbnb API python" in practice means calling this actor: one POST with a place or listing ids and a stay, a dataset of typed rows back. The same call from JavaScript, curl or an HTTP node is an Airbnb calendar API when calendarMonths is set. The examples are complete; the row schemas are documented above and in the Output tab, so an agent or a script can filter on price_nightly_effective, occupancy_90 or available without parsing HTML.

from apify_client import ApifyClient
client = ApifyClient("<TOKEN>")
run = client.actor("kestrel/airbnb-prices-calendar").call(run_input={"locationQueries": ["Lisbon, Portugal"], "checkIn": "30 days", "checkOut": "33 days", "adults": 2, "currency": "EUR", "maxListingsPerQuery": 80, "calendarMonths": 3})
rows = list(client.dataset(run["defaultDatasetId"]).iterate_items())
listings = [r for r in rows if r["type"] == "listing"]; cal = {r["id"]: r for r in rows if r["type"] == "calendar"}
for l in sorted(listings, key=lambda r: r["price_nightly_effective"] or 0)[:10]: print(l["name"], l["price_nightly_effective"], "occupancy 90d", cal.get(l["id"], {}).get("occupancy_90"))
import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: '<TOKEN>' });
const run = await client.actor('kestrel/airbnb-prices-calendar').call({ listingIds: ['26119973'], checkIn: '30 days', checkOut: '33 days', includeDetails: true, calendarMonths: 12 });
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items.filter(r => r.type === 'calendar').map(r => [r.id, r.occupancy_90, r.next_available]));
curl -X POST "https://api.apify.com/v2/acts/kestrel~airbnb-prices-calendar/run-sync-get-dataset-items?token=<TOKEN>" -H "Content-Type: application/json" \
-d '{"locationQueries": ["Lisbon, Portugal"], "checkIn": "2027-03-12", "checkOut": "2027-03-15", "maxListingsPerQuery": 20}'
  • n8n, Make, Zapier: run on a schedule, read the dataset, alert on price drops or occupancy changes.
  • AI agents and MCP: available through the Apify MCP server and Apify AI; the section "Airbnb scraper for AI agents" above maps each request to its input, and rows are flat and typed so "cheapest entire homes in Lisbon for these dates with a rating over 4.8" is a filter.
  • Google Sheets, BigQuery: export the dataset from the run or through Apify integrations.

The rows are public listing data as any logged-out visitor sees it: prices, availability, ratings, amenities, and reviews as published. No login, no personal data beyond what is shown publicly (reviewer first names, host first name); it does not scrape Airbnb host emails or phone numbers. Terms of service and local law vary — check that your use complies, and use the data for your own analysis.

Limits and honest notes

  • One Airbnb search stops at about 280 listings. Whole-area mode (maxListingsPerQuery above 280) partitions the search by price, then room type, then (map areas) quadrants, until each slice is at most 270 by Airbnb's own count; a slice that still holds more (thousands of listings at one price and room type on a place search) is read to its 280 and counted in truncated. Airbnb re-ranks every request, so a listing can surface twice: it is delivered and billed once.
  • A listing Airbnb does not show publicly (private, unlisted, snoozed or removed) gets a free not_public status row and the run's other listings are delivered as usual; a run whose every listing is like that ends green as an input problem, with nothing charged.
  • occupancy_* counts nights that are not available; blocked nights and booked nights look the same in the calendar.
  • Prices are what a logged-out visitor sees in currency for the stay and guests given, always before taxes (see "Airbnb prices for exact dates"); price_taxes is filled only when the answer came from an exit Airbnb shows taxes to (EU, UK), so for the taxes on every row set proxyConfiguration to Apify residential proxies of such a country. Airbnb does not offer every currency: for one it does not (GEL, 2026-11-18) it prices in USD, and those rows arrive without amounts and free, with currency_mismatch saying so.
  • Check-in and checkout times and the cancellation policy left the listing page's own answer in November 2026 (run cG9uP9fRpy3q0IkFL: 0 of 10); since version 1.15 they come from the listing's "Things to know" answer, one more small request per detailed listing. If Airbnb ever retires that query, those fields stay empty, the run says so and the paid rows are delivered as usual.
  • Airbnb rotates the hashes of its persisted queries every few weeks; the actor picks up a rotation by itself from the site's bundles, including the reviews query that only loads on demand. When Airbnb retires or renames a query (the calendar query was renamed on 2026-11-14 and its variables changed), the run does not guess: it fails with a message naming the retired query and the similar ones the site registers now, after delivering and billing every row that did not depend on it. A daily canary run watches all of it.

FAQ

Does it need an Airbnb API key, login or cookies?

No. It uses the public key every visitor's browser uses and no session; there is no API key for guests to request.

Is there a free Airbnb scraper here?

Apify's free plan covers small runs, and the actor never charges for unknown places, unpriced listings or empty searches. The Python and JavaScript examples above are complete.

Why does a place come back as unknown_area?

An unrecognised place is answered with listings from all over the world; the actor detects that (Airbnb gives such a search no canonical URL), returns a status row and charges nothing. A country such as "India" is a known place, however far apart its listings are. Use the spelling the site suggests, e.g. "Lisbon, Portugal".

Can I get more than 280 listings for a city?

Yes: set maxListingsPerQuery up to 10,000. The run splits the search into price bands (then room types, and boxes for a map area) that Airbnb answers in full, de-duplicates them and delivers every listing once — Lisbon held about 6,100 for one December stay, Porto 4,888. SUMMARY coverage and the status row give Airbnb's own count next to what was delivered.

Is occupancy the same as bookings?

No. It is the share of nights not available; hosts also block nights. Over a market it tracks demand well; for one listing treat it as an upper bound.

Can I track Airbnb prices and availability over time?

Yes. Schedule the actor and save each run's dated rows (fetched_at) externally; each run has a separate dataset. Relative dates observe a moving stay at the same lead time; fixed dates observe the same stay. For new listings only, see "Monitor new Airbnb listings" (onlyNew). Calendar summaries report unavailability, including both blocked and booked nights.

Can I scrape Airbnb host emails or phone numbers?

No, by design: the rows carry the host's public profile (with includeHostDetails: name, Superhost, verification, languages, years hosting, number of listings), nothing off-platform.

Can I download an Airbnb dataset as CSV?

Every run's dataset downloads as JSON, CSV or Excel from the run page or the API — an Airbnb dataset for the market, dates and listings you chose, with calendar days as nested arrays. For one CSV row per listing and day, add unwind=days: https://api.apify.com/v2/datasets/<datasetId>/items?format=csv&unwind=days&fields=type,id,name,days.

Can I get Airbnb listing data without dates?

Yes: leave checkIn/checkOut empty and each listing row carries the default displayed price and price_nights; add includeDetails for description, amenities and host.

Why is a listing id reported as rounded_id?

It was sent as a JSON number above 2^53 (Airbnb's newer ids have 18–19 digits), and JSON rounds such numbers before the run starts, so it may name another listing. It is not fetched and costs nothing: send listing ids as strings, ["1663639623527062900"], or as /rooms/ URLs.

Which languages and currencies?

Any site locale for text (locale) and any ISO currency for prices (currency); both are echoed in every row.

Last verified working: 2026-11-18.

Ready-made runs

Each link opens this actor with the input already filled in — press Start and it runs. Every one is capped, so what it bills is on the page before you run it. What each returns is spelled out at mtedj.github.io/kestrel-actors-examples/recipes.html.

Prices and availability are one half of a listing's story; reviews are the other. These share the same row shape and billing model:

All of them bill per delivered row, never charge for rows a filter or a spending limit removed, and write an Apify dataset you can export to CSV, Excel or JSON.