Booking.com Price Scraper & Room Rate Tracker

🏨 What this Actor does

Track Booking.com hotel prices room by room: one row per room type, per night, for every property you give it, with the price, what that price covers, the booking terms, and whether the room can still be booked.

Put your hotel and its competitors in one run and it also ranks you against them night by night, remembers yesterday, and can email you the morning read.

A value Booking.com does not show stays empty (null); it is never filled with a guess. Ranks, medians and labels are computed from what was read, and the field reference says how.

Used this Actor before September 2026? Nothing you use was removed. See what changed.

---

⚡ Quick start

Start small, see the output, then grow it.

1. Paste this input and run it. Three Perth motels, three nights, one adult: about 50 rows in a few minutes.
2. Swap in your own hotels. Your hotel and its competitors in Hotel URLs, and your hotel again in My hotel URL. 7–14 nights is a good daily window.
3. Schedule it daily, at the same hour. Use Actor → Schedule, or the Schedule this run bar above your results. The change columns and the competitor moves compare each run with the previous one, so the first run has nothing to compare with.

{
  "hotelUrls": [
    { "url": "https://www.booking.com/hotel/au/flag-motor-lodge.html" },
    { "url": "https://www.booking.com/hotel/au/great-eastern-motor-lodge.html" },
    { "url": "https://www.booking.com/hotel/au/ibis-budget-perth-airport.html" }
  ],
  "myHotelUrl": "https://www.booking.com/hotel/au/flag-motor-lodge.html",
  "daysAhead": 3,
  "adults": 1,
  "rooms": 1,
  "currency": "AUD"
}

⚠️ Put your hotel and its competitors in the same run. The ranking compares the properties inside one run. Scrape them one at a time and every rank, median and gap comes back empty.

---

⭐ Start here: the 12 fields that matter most

Every example below comes from a real run on four Perth airport motels, 25 September 2026, one adult, prices in AUD.

Field	What it tells you	Example
nightly_total_price	The price Booking.com shows for the rate that fits your party. Compare on this one, and read tax_clarity_label beside it.	217 (rounded from 217.27; with the 21.73 reported tax on top: 239.00)
lowest_bookable_rate	The cheapest rate your party can book for this room, whatever its terms: the price a guest sees first.	213, beside a selected 235 on the same room
tax_clarity_label	Whether that price is the final cost. Only ✅ means final.	⚠️ Additional taxes reported
availability_label	Availability at a glance, with a colour.	🔴 Only 2 rooms left · ⚫ Sold out
rooms_left + stock_confidence	Rooms left, only when Booking.com states it, and how far to trust that number.	2 + exact · null + hidden
booking_terms_summary	Cancellation, breakfast and payment in one line.	Free cancellation · Breakfast not included · Pay now
rate_selection_reason	Why this rate was picked, and which cheaper one was not.	…a cheaper non-refundable rate (180) was available but not selected…
price_move_label	What the price did since the last run.	▲ +191 (+82.68%)
is_my_hotel	true on every row of your own property.	true
lead_in_rate (Market by date)	A property's cheapest bookable rate that night. The ranking is built on it.	217
compset_rank / gap_to_market_median (Market by date)	Where the property sits that night: 1 = cheapest.	2 of 4 · -50
demand_pressure (Market by date)	How many of the property's room types Booking.com shows unavailable that night, as a level.	high — "5 of 6 room types unavailable"

---

📬 The daily digest

The morning read of a run, built only from that run's own figures. This is the real one from the Perth run of 25 September, shortened:

Booking.com rate shop · Fri 25 Sep
4 properties · 10 nights (Fri 25 Sep – Sun 04 Oct) · 1 adult · AUD
Prices are before tax, as Booking shows them from US: the guest pays the tax on top.

👉 Wed 30 Sep: you are 100 under your competitors' median (217 vs 317).

YOUR HOTEL — Flag Motor Lodge
• Cheapest on 0 of 4 ranked nights; most expensive on 1.
• Above the market Sat 03 Oct: 199 vs competitors' median 181 (+18).
• Below the market Wed 30 Sep: 217 vs competitors' median 317 (−100).
• You are sold out: Fri 25 Sep, Sat 26 Sep, Sun 27 Sep, Mon 28 Sep, Tue 29 Sep, Thu 01 Oct.

NEXT NIGHTS — lead-in (cheapest bookable rate)
• Mon 28 Sep · cheapest 180 (ibis Budget Perth Airport) · competitors' median 272 · you: sold out · 1 of 4 sold out
• Wed 30 Sep · cheapest 180 (ibis Budget Perth Airport) · competitors' median 317 · you 217 (#2 of 4)

NIGHTS TO WATCH — the market is filling
• Sun 27 Sep: 3 of 3 competitors sold out

COMPETITOR MOVES SINCE THE LAST RUN
🟢 Great Eastern Motor Lodge · Fri 25 Sep: Superior Double or Twin Suite back on sale at 317 (249 before it sold out)
⚫ Sanno Marracoonda Perth Airport Hotel · Fri 25 Sep: King Room sold out

Run: ✅ 40 of 40 pages, 240 rows.

Part	What it says
Header	The hotel's own date, how many properties and nights, the party and the currency. A line under it says when the prices are before tax.
👉 The one line	The strongest signal of the run: a night where most competitors are sold out while you still sell, otherwise the night you sit furthest under your competitors, otherwise furthest over.
Your hotel	Your position against your competitors' median, your own rate left out: a median that includes you moves when you move. It needs two competitor rates.
Next nights	The next 7 nights: the cheapest property, the competitors' median, your rate and rank, and how many are sold out. A night with a property that could not be read says so before its numbers.
Nights to watch	Nights where half or more of your competitors have nothing left to sell.
Competitor moves	Price moves, sell-outs and rooms back on sale since the last run. A room back on sale is shown as a return, with the price it had before it sold out.
⚠️ Run	Only when pages failed: how many, which hotel and which nights, and what to do.

Getting it

• By email: fill digestEmail. On plans where Apify refuses to start its Send Email Actor from inside a run, the log says "Digest email NOT sent".
• Anywhere, on any plan: put an n8n, Make or Zapier webhook in digestWebhookUrl and send the email from there. The JSON carries the text and every figure in structured form.
• Always: the digest is also saved with the run, as DIGEST (text) and DIGEST_DATA (JSON), and linked on the Output tab.

---

💰 Pricing

You pay per result: $3.50 to $5.00 per 1,000 rows, depending on your Apify plan, plus Apify platform usage. Every row is a result, sold-out rows included.

The Perth example, 4 motels × 10 nights, wrote 240 rows: $0.84 to $1.20 in result fees for the run, platform usage not included. The Pricing tab shows the price for your plan. To pay for fewer rows:

• price fewer nights (daysAhead);
• turn on tieredCadence for long horizons;
• turn off includeSoldOut, knowing you lose the demand read.

---

📦 What you get

Output	Where to find it	One record per	Use it for
Room rows	Output tab: Rate shopping, Rate shopping V2, Full detail, CSV, Excel	room type × night × proxy country	Every price, term and availability, room by room
Market by date	Output tab: Market by date (market_summary dataset)	property × night	Your rank, the median, the gaps, and the demand read
Daily digest	Email or webhook, and DIGEST (text) / DIGEST_DATA (JSON) in the run's key-value store	run	The morning read: one screen of text
Hotel metadata	hotel_metadata dataset	property × run	The full hotel snapshot: reviews, description, every facility
Run report	RUN_REPORT in the run's key-value store, and the run's status line	run	What worked, what did not, and whose fault it was

The room rows come in four views of the same data:

View	Fields	For	Through the API
Rate shopping	25	The working screen: price, tax, terms, availability	?view=rate_shopping
Rate shopping V2	52	The daily screen: adds the movement since the last run, the lead-in rate and how many rooms are still bookable	?view=rate_shopping_v2
AI input (compact)	41	Handing the data to an AI model	?view=ai_input
Full detail	104	Every field, in reading order	?view=full_detail
Room comparison (legacy)	27	Kept so API calls made before the update keep working	?view=room_comparison
AI room picker (legacy)	47	Kept so API calls made before the update keep working	?view=ai_room_picker

Add the view to the dataset's items URL, for example …/datasets/<id>/items?view=rate_shopping_v2. The Download as CSV / Excel links on the Output tab carry every field, with the rate plans flattened into columns. A download made from a view carries that view's columns only.

---

🔍 How to read the data

• null is never 0. An empty price, count or tax means Booking.com did not state it. It is never filled with a guess.
• The price is what Booking.com shows from the proxy country. Booking.com shows US visitors the price before tax, with the tax on top. Elsewhere it depends on the country and the property. On the Perth runs every rate was 10 % under its total with the reported tax: 217.27 (shown as 217) + 21.73 = 239.00. tax_clarity_label on each row and taxes_on_top on each night say which one you are looking at.
• Booking.com prices by visitor country. The same room can cost more or less for a French visitor than for a US one. Compare rows from the same proxy_country, with the same currency. To see prices closer to what local guests see, load the pages from the hotel's own country, in its currency.
• One night per row. Multi-night stays are not quoted.
• Genius rates are flagged, not hidden. is_genius_rate tells you when a price is a member rate.
• Sold out means not bookable on Booking.com for this search: this party, this one night. A room type can be closed on Booking.com, held for other channels, or need a longer stay. It is not the hotel's occupancy, and demand_pressure is a label read from it.
• A sold-out room still has a row, marked ⚫, so a filling date never looks cheaper than it is. With includeSoldOut off, those rows and the demand read disappear.
• Two runs minutes apart can differ. Booking.com tests prices and layouts on its visitors. Every row carries its capture time and country.

---

📅 Scheduling and history

The change fields and the competitor moves compare each run with the previous one. The previous prices live in a named key-value store, booking-price-history, in your Apify Storage.

• Do not delete that store. It is the memory. Delete it and every row reads ✦ New again until the history rebuilds. Deleting a run's dataset changes nothing: the next run never reads it.
• Run at the same hour every day. A second run on the same day compares with the earlier one, so it shows only a few hours of moves.
• A room-night is known by hotel + night + room + party + currency + proxy country. Change the party, the currency or historyStoreName and every row starts again as ✦ New. Adding a hotel or a country starts only that one. Changing daysAhead, startDate, concurrency or the toggles changes nothing.
• A page that fails keeps its older baseline. The next run compares with the last time it was read, and days_since_previous_capture says how long ago that was.
• The store cleans itself. An entry is dropped once its night has passed, or after 30 days without being seen.

---

📥 Input

One of hotelUrls, searchUrl or searchText is required.

What to scrape

Field	Default	What it does
hotelUrls	—	Booking.com hotel pages. The plain URL is enough; dates in the URL are ignored.
myHotelUrl	—	Your own property, also kept in hotelUrls. Every row of yours gets is_my_hotel: true and the digest ranks you.
searchUrl	—	A Booking.com search-results URL with your filters. Every hotel found is scraped.
searchText	—	A free-text destination, such as riad with pool in fes. Used only when the two above are empty.
maxProperties	50	The most hotels to take from a search.
propertyType	—	Filter a searchText search by type.
minReviewScore	0	Skip hotels below this score. 0 = no limit.
maxReviewScore	0	Skip hotels above this score. 0 = no limit.

Dates and party

Field	Default	What it does
startDate	today	First check-in, YYYY-MM-DD. Empty = today at each hotel. Leave it empty on a schedule: a fixed date slides into the past.
daysAhead	3	How many nights to price (1–365). Each night is one page per hotel, so this drives time and cost.
adults	2	Adults in the party. It decides which rate is selected.
children	0	Children in the party.
childrenAges	—	One age per child (0–17). Booking.com prices a family by age.
rooms	1	Rooms. Capacity is checked per room.
currency	USD	One of 23 currencies, the same for every price.

What to keep

Field	Default	What it does
includeSoldOut	true	Keep a row for every sold-out room. Needed for the demand read. Every row is a result you pay for, and on busy dates most rows can be sold out.
trackPriceChanges	true	Compare with the last run and fill the change fields. No extra requests.
historyStoreName	booking-price-history	Leave it as it is. Give two jobs different names only if their history must not mix. Renaming starts the history over.
includeLosAnalysis	false	Adds los_prices and has_los_discount (booking several of a room at once). No extra page loads.
includeWholesaleRates	false	Also capture partner and wholesale rates, tagged apart from retail.
tieredCadence	false	Near nights daily, mid nights every 2 days, far nights every 3 days. About half the cost for long horizons.

Digest and alerts

Field	Default	What it does
digestEmail	—	Email the daily digest, through Apify's Send Email Actor. Some Apify plans cannot start it: use the webhook below instead.
digestWebhookUrl	—	Slack or Discord webhook: the digest as a message. Any other URL (n8n, Make, Zapier): JSON {"text": …, "digest": {…}}.
webhookUrl	—	Slack or Discord webhook for error alerts.

Proxy and speed

Field	Default	What it does
proxyConfiguration	Residential	Your choice of Apify proxy: residential or datacenter. Datacenter costs less, but Booking.com blocks it sooner: some nights can come back empty, or the run can fail.
proxyCountries	["US"]	The countries to load the pages from. Several countries = the same hotels once per country, tagged with proxy_country.
useApifyProxy	true	Old field, hidden from the form. proxyConfiguration is read first.
concurrency	3	Hotels scraped at once (1–5), each in its own browser. Lower it on a small Apify plan: fewer browsers use less memory.
delayBetweenRequests	3	Seconds between two nights of the same hotel.

---

🧯 When a run goes wrong

Every run writes a report, RUN_REPORT, into its key-value store, and its headline becomes the run's status line.

verdict	Meaning
complete	Every page returned data.
partial	Some nights failed. Every row you got is real. The report lists the hotels and nights to re-run, and the digest names them.
failed	Nothing came back. The headline names the cause and what_to_do gives the fix.
crashed / nothing_attempted	The run stopped before scraping, or the input had nothing to scrape.

A row that could not be read says why in availability_label, and error_type says whose failure it was: proxy, page or browser.

---

🤖 Use it from Claude or ChatGPT

With the Apify MCP server, an AI assistant can run this Actor and read its results. Add noraview/booking-price-scraper to the MCP server's tools in your AI client, with your own Apify API token, and keep that token private.

When you hand it the results, point it at the AI input (compact) view, and tell it: my hotel is the row where is_my_hotel is true.

---

📖 Field reference

Every field, grouped by the question it answers. Empty when… says when a field is null and why. null always means not known, never 0, never "no".

Room rows: which room, which night

Field	What it tells you	Example	Empty when…
hotel_name_clean	The property's name, without the extras Booking.com puts in its page title.	Flag Motor Lodge	Never
is_my_hotel	true on every row of the property you gave as myHotelUrl, false on the others.	true	myHotelUrl was left empty. Not yours and you never said which one is yours are different answers
check_in_date	The night this row prices (YYYY-MM-DD).	2026-09-30	Never
check_out_date	The morning after. Every row prices one night.	2026-10-01	Never
lead_time_days	Nights between the run and arrival, counted from the hotel's own date. 0 = tonight. Segment on this before comparing anything.	5	The date could not be read
day_of_week	Arrival weekday.	Wed	The date could not be read
is_weekend	Arrival on a Friday or Saturday, the hotel weekend. Use day_of_week where your market's weekend differs.	false	The date could not be read
room_name_clean	The room's name, without the bed phrase and deal tags, so the same room reads the same across nights.	Standard Room	Never
room_id	Booking.com's id for the room type. Stable across runs.	117422001	Booking.com did not expose it

Room rows: the price ⭐

Field	What it tells you	Example	Empty when…
nightly_total_price ⭐	The price Booking.com shows for the selected rate, as a guest sees it (whole number). Tax that Booking.com puts inside the price is included; tax it charges on top is not. Read tax_clarity_label beside it.	217	Sold out, or no rate fits your party (availability_label says which)
nightly_total_price_exact	The same rate, unrounded, from Booking.com's own data. Use it when cents matter.	217.27	Same as above
lowest_bookable_rate ⭐	The cheapest rate your party can book for this room tonight, whatever its terms. Same tax basis as nightly_total_price. They differ when a cheaper rate is non-refundable or is priced for more guests. Rates too small for your party are never counted.	213 (Sanno, King Room, Fri 02 Oct, where the selected rate was 235)	Sold out, or no rate fits your party
lowest_bookable_terms	The terms of that cheapest rate.	Non-refundable · Breakfast not included · Pay now	Same as above
nightly_base_price	The room rate without tax. Only filled when Booking.com states the tax inside the price; never worked out from an assumed rate.	null (tax charged on top in these runs)	Booking.com states no included tax
taxes_and_fees_total	All the tax Booking.com reports for the night: inside the price plus on top.	21.73	Booking.com states no tax
tax_clarity_label ⭐	Whether the price is final. ✅ Taxes included: every reported charge is inside the price. ⚠️ Part included, more on top: add taxes_excluded for the real cost. ⚠️ Additional taxes reported: tax is charged on top and the amount is known. ⚠️ Final cost not confirmed: no breakdown was stated.	⚠️ Additional taxes reported	Sold out
primary_tax_label	The main tax line of the selected rate, in Booking.com's words.	10 % Tax	No tax line on the rate
total_stay_price	nightly_total_price × number_of_nights. With one night per row, the same number.	217	Same as nightly_total_price
currency	The currency of every price on the row.	AUD	Never
price_basis	What the price fields represent, as a short machine-readable token. The guest count is the selected rate's capacity (max_guests), not your party; for the tax, read tax_clarity_label.	1_night_2_guests_taxes_not_confirmed · sold_out	Never
is_genius_rate	true when the price is a Genius (member) rate, not the public price. Never compare your public rate with a competitor's Genius rate.	false	No rate selected
original_price	The struck-through price, when Booking.com shows one on the room.	null	No strikethrough shown
discount_percent	The discount against original_price, computed on the legacy price. Do not recompute it against nightly_total_price.	null	No strikethrough shown

Room rows: why this rate

A room usually has several rate plans. The row reports one, picked by fixed rules: it must fit your party, a refundable rate beats a cheaper non-refundable one, and among equals the cheapest wins.

Field	What it tells you	Example	Empty when…
rate_selection_reason ⭐	Why this rate was picked, in plain words, including the cheaper rate it did not take. Written by the same rules that pick the rate; never by an AI.	Selected the refundable rate compatible with 1 guest(s) per room; a cheaper non-refundable rate (180) was available but not selected, since a refundable rate at the same capacity is preferred.	No rate could be selected
occupancy_match_status	How the rate was matched to your party: capacity_compatible (it fits), uncertain_children (children in the party, whose counting Booking.com does not state), no_compatible_rate (nothing fits).	capacity_compatible	Never
max_guests	The capacity of the selected rate: what it will take, not what you asked for.	2	No rate fits your party
required_guests_per_room	Your party split across your rooms: the capacity a rate must have to count.	1	Never
rate_options_count	How many rate plans Booking.com offers for this room tonight. 1 = the room is not fenced at all.	1	Never (0 when sold out)
rate_spread	Dearest minus cheapest rate plan for this room: how widely the property fences its rates. 0 = several plans at the same price.	null (one plan only)	The room has only one rate plan
refundable_premium	Cheapest refundable rate minus cheapest non-refundable rate: what flexibility costs here.	null	The room does not offer both
primary_rate_option	The selected rate as one object. Every value in it comes from the same Booking.com rate: price, public price, capacity, cancellation, free-cancellation date, meal plan, breakfast, pay later, Genius, wholesale, taxes and their breakdown.	{ "price": 217, "max_guests": 2, … }	Sold out, or no rate fits your party
benchmark_rate_option	The same object for a fixed 2-adult rate, whatever your search. For benchmarking only; not for "which room should this guest book".	{ … }	Same as above
rate_options	Every rate plan found for the room, in the page's raw shape, to inspect the alternatives. Also carries each plan's strikethrough price and tax.	[ { … }, { … } ]	Empty list when sold out

Room rows: the booking terms

All of these describe the selected rate, the same one nightly_total_price comes from.

Field	What it tells you	Example	Empty when…
booking_terms_summary ⭐	Cancellation, breakfast and payment in one line. A condition that is not known is left out, never written as "No". Breakfast extra (the page sells it) is told apart from Breakfast not included.	Free cancellation · Breakfast not included · Pay now	No rate selected
cancellation_policy	fully_refundable, partially_refundable, non_refundable or unknown.	fully_refundable	No rate selected
is_refundable	true only for fully_refundable. A partial refund is not free cancellation.	true	No rate selected
free_cancellation_until	The date in "Free cancellation before …": cancel before it and the guest pays nothing.	2026-09-29	Non-refundable, or no date printed
free_cancellation_days_before_arrival	The same deadline as days before arrival: the policy itself, comparable across properties and dates.	1	No free-cancellation date
meal_plan	The meal line of the rate, in Booking.com's words.	Good breakfast AUD 12	The rate carries no meal line
is_breakfast_included	Whether breakfast is in the price.	false	No rate selected
book_now_pay_later	Whether the guest can book now and pay later.	false	No rate selected
min_stay	Minimum nights for the rate.	null	Booking.com does not show it on the property page

Room rows: availability and stock

Field	What it tells you	Example	Empty when…
is_sold_out	Booking.com's own sold-out flag for this room and night.	false	Never
availability_status	available, low_availability (4–5 left), low_availability_urgent (1–3 left) or sold_out. It only moves on a count Booking.com states, so it reads available whenever no count is shown.	available	Never
availability_label ⭐	The same fact for a human: a colour and words. Also says when a room is on sale but has no rate for your party, and why a row has no data.	🔴 Only 2 rooms left · ⚫ Sold out · ⚪ No rate for 1 guest · ⚠️ Not scraped — the page could not be read	Never
rooms_left ⭐	Rooms left, only when Booking.com prints "Only N left". null means unknown, not plenty. 0 on sold-out rows.	2	On sale and no count shown (the usual case)
stock_confidence ⭐	How far to trust rooms_left: exact (Booking.com's own badge), hidden (on sale, no count shown), inferred (sold-out rows), capped (10, Booking.com's display ceiling). On the Perth run: 15 exact, 50 hidden, 175 inferred.	exact	Never
bookable_now_label	How many of this room a guest can still book, said no stronger than the page proves. 4 left: Booking.com's badge, or a quantity list that stops below another room's. 10+: at least that many.	2 left · 10+	Sold out, or no quantity list on the page
max_rooms_per_booking	The top of this room's quantity list: the most a guest can put in one booking now. It can be the property's own ceiling rather than the stock; bookable_now_label says which.	8	No quantity list on the page
max_nights_bookable	Booking.com's own count of this room bookable at once, from the page's rate data. Rooms, not nights: the name is historical. Prefer max_rooms_per_booking and bookable_now_label.	2	Not in the page's rate data
los_prices	Booking.com's price for booking 1, 2, 3… of this room at once, for this night (the quantity list). Only with includeLosAnalysis. Not length-of-stay prices.	{"1": 130, "2": 260, …}	includeLosAnalysis is off
has_los_discount	true when booking several of this room costs less per room than one. Only with includeLosAnalysis.	false	includeLosAnalysis is off

Room rows: what changed since the last run

These compare each room-night with the previous run of the same job. On the first run there is nothing to compare, so they read ✦ New or null. All of them are empty when trackPriceChanges is off.

Field	What it tells you	Example	Empty when…
price_move_label ⭐	The move at a glance: an arrow and its size.	▲ +191 (+82.68%) · – Unchanged · ✦ New	Sold out now or at the last run: no price to compare
price_direction	The same as a token: up, down, unchanged, or new the first time a room-night is seen, sold out or not.	up	Sold out now or at the last run, after its first sighting
price_change	Today's nightly_total_price minus the last one.	191	First sighting (never 0, which would say the price held)
price_change_pct	The change as a percentage of the last price.	82.68	Same as above
previous_price	The price at the last capture.	231	First sighting
previous_captured_at	When that last capture was.	2026-09-24	First sighting
days_since_previous_capture	Days since that capture. Read it before price_change: a move over five days is not a daily move. 0 on a second run the same day.	1	First sighting
days_since_price_change	How long the price had held up to this run. A competitor that has not moved in twelve days is a finding.	3	First sighting
first_seen_price	The first price ever recorded for this room-night.	217	Never priced yet
first_seen_at	When that first price was recorded.	2026-09-22	Never priced yet
availability_change	new, unchanged, became_sold_out or came_back. A room selling out often moves before any price does.	came_back	Price tracking off, or error rows
days_sold_out	How long the room has been off sale. On the run it comes back, how long it was gone. On a come-back row the price is compared with the last price before the sell-out, so the baseline is this many days older.	3	On sale and never sold out
sold_out_episodes	How many separate times this room-night has gone off sale. 3 or more: it keeps going on and off sale.	1	Price tracking off

Room rows: the room

These are read from the room's own listing. Sold-out rows carry fewer of them.

Field	What it tells you	Example	Empty when…
room_name	The room's name exactly as Booking.com shows it.	Standard Room	Never
bed_type	The beds, as Booking.com lists them.	1 extra-large double bed	No bed line for the room
room_capacity	How many the room takes. When the page does not say, the largest capacity a rate proves: read it as at least.	2	Not stated and no rate proves it
room_size_sqm	Size in square metres.	16	Not stated
room_view	The view.	Garden view	Not stated
has_private_bathroom	Private or ensuite bathroom.	true	Not stated
room_amenities	The amenities that change the price: balcony, minibar, spa bath…	["air conditioning"]	Empty list when none
facilities	The room's full facility list, often 30–40 items.	["Air conditioning", "Ensuite bathroom", "Flat-screen TV", "Free WiFi", …]	Empty list when none
photo_count	How many photos Booking.com holds for this room type.	7	The page exposes none (never a false 0)
room_photo_url	Booking.com's first photo of the room, at the size Booking.com serves. No image is downloaded.	https://cf.bstatic.com/xdata/images/hotel/max1024x768/116836127.jpg?k=…	No photo exposed

Room rows: the hotel

The same on every row of a property. The full snapshot is in the Hotel metadata dataset.

Field	What it tells you	Example	Empty when…
hotel_name	The name as Booking.com's page title gives it. Use hotel_name_clean for grouping.	Flag Motor Lodge (Motel) (Australia) deals	Never
hotel_url	The property page that was scraped.	https://www.booking.com/hotel/au/flag-motor-lodge.html	Never
booking_hotel_id	Booking.com's id for the property.	1174220	Not exposed
hotel_review_score	Review score, 0–10.	7.5	Not shown
hotel_review_count	Reviews behind that score.	4265	Not shown
hotel_star_rating	Official star rating.	3	No stars shown
hotel_neighborhood	The area, from Booking.com's location data.	Rivervale	Booking.com gives no area
hotel_address	The full address.	129 Great Eastern Highway, Rivervale, 6103 Perth, Australia	Not shown
property_type	Hotel, motel, riad, apartment…	Motel	Not stated
hotel_latitude	Latitude.	-31.9526	Not exposed
hotel_longitude	Longitude. Also used to find the hotel's own "today".	115.9159	Not exposed
hotel_checkin_from	Check-in start time as Booking.com states it.	2:00 PM	Not shown when the page is read
hotel_sustainability_certified	Booking.com's Travel Sustainable badge.	true	Hotel details could not be read

Room rows: the search and the run

Field	What it tells you	Example	Empty when…
requested_adults	The adults this row was searched for.	1	Never
requested_children	The children this row was searched for.	0	Never
requested_children_ages	The children's ages sent to Booking.com.	[4, 9]	No ages given
requested_rooms	The rooms this row was searched for.	1	Never
number_of_nights	Nights covered by the row. Always 1.	1	Never
proxy_country	The country the page was loaded from. Booking.com prices by visitor country, so compare rows from the same country.	US	Never
scrape_timestamp	When the run started (UTC).	2026-09-25T12:59:15Z	Never
sort_key	hotel · date · room · room id, lowercased. Sort on it and any two exports line up row for row.	flag motor lodge|2026-09-30|standard room|117422001	Never

Market by date: one row per property, per night

Built from the room rows of the same run, with no extra requests. With a single property in the run, every comparison field is null: rank 1 of 1 is not information.

The examples are Flag Motor Lodge on Wed 30 Sep, from the same Perth run.

Which property, which night

Field	What it tells you	Example	Empty when…
record_type	Always market_summary.	market_summary	Never
hotel_name_clean	The property.	Flag Motor Lodge	Never
hotel_url	Its page.	https://www.booking.com/hotel/au/flag-motor-lodge.html	Never
booking_hotel_id	Booking.com's id for it.	1174220	Not exposed
is_my_hotel	true for your property.	true	myHotelUrl empty
check_in_date	The night.	2026-09-30	Never
check_out_date	The morning after.	2026-10-01	Never
lead_time_days	Nights between the run and arrival.	5	Date unreadable
day_of_week	Arrival weekday.	Wed	Date unreadable
is_weekend	Friday or Saturday arrival.	false	Date unreadable
requested_adults	The party searched.	1	Never
requested_children	Children searched.	0	Never
requested_rooms	Rooms searched.	1	Never
proxy_country	The country the prices were seen from.	US	Never
currency	The currency of every price.	AUD	Never
scrape_timestamp	When the run started (UTC).	2026-09-25T12:59:15Z	Never
sort_key	For stable sorting across exports.	flag motor lodge|2026-09-30||	Never

The property's own night

Field	What it tells you	Example	Empty when…
lead_in_rate ⭐	The cheapest rate your party can book at the property that night, across its rooms and whatever the terms. The ranking is built on it.	217	Nothing bookable
median_rate	The median nightly_total_price of its rooms on sale.	217	Nothing bookable
highest_rate	Its most expensive room on sale.	217	Nothing bookable
rate_range	highest_rate − lead_in_rate: how far it spreads its rooms that night.	null (one room on sale)	Fewer than two rooms on sale
refundable_premium_median	The median cost of flexibility across its rooms.	null	No room offers both
taxes_on_top	true when every rate on sale adds tax on top of the price shown, false when every rate has it inside. Compare properties on the same basis.	true	The rates disagree or do not say

Demand

Field	What it tells you	Example	Empty when…
room_types_offered	Room types the property lists at all.	6	Sold-out tracking off
room_types_available	Room types bookable that night for your party.	1	Never
room_types_sold_out	Room types listed but not bookable.	5	Sold-out tracking off
sold_out_ratio	Sold out ÷ offered: the share of the property's room types Booking.com shows unavailable for this search. Not occupancy.	0.833	Sold-out tracking off
rooms_length_restricted	Room types on sale that are down to their last room. The name is historical.	0	Never
restricted_ratio	That count ÷ rooms on sale. Reported, but does not move demand_pressure.	0.0	Nothing on sale
demand_pressure ⭐	A label read from sold_out_ratio: sold_out, high (half or more of the room types unavailable), building (some) or normal (none). It describes availability on Booking.com, not occupancy.	high	Sold-out tracking off
demand_pressure_reason	The numbers behind the level, in words.	5 of 6 room types unavailable	No room types at all

Against the others that night

Field	What it tells you	Example	Empty when…
compset_size	Properties with a bookable rate that night. The comparison is only this wide.	4	Only one property in the run
compset_rank ⭐	This property's place by lead_in_rate, 1 = cheapest. A sold-out property takes no rank.	2	Sold out, or no one to compare with
market_lead_in_rate	The cheapest lead_in_rate that night.	180	No one to compare with
market_median_rate	The median lead_in_rate of every property that night, yours included. The daily digest compares you with your competitors only (317 that night).	267	No one to compare with
gap_to_market_median ⭐	This property minus that median. Positive = above the market.	-50	Sold out, or no one to compare with
gap_to_cheapest	This property minus the cheapest.	37	Same as above
price_position	cheapest, below_median, at_median, above_median, most_expensive, or sold_out.	below_median	No one to compare with

Coverage: did every property load?

Field	What it tells you	Example	Empty when…
compset_expected	Properties the run asked for.	4	Never
compset_observed	Properties that produced data that night. Fewer means one failed, and its absence moves the median.	4	Never
compset_coverage	Observed ÷ expected. Below 1.0, every market figure on the row is built on an incomplete set.	1.0	Never
coverage_warning	A sentence naming how many properties are missing.	null	The set is complete

Hotel metadata: one row per property, per run

Field	What it tells you	Example
record_type	Always discovery.	discovery
hotel_url	The property page.	https://www.booking.com/hotel/au/flag-motor-lodge.html
hotel_name	Its name, as the page title gives it.	Flag Motor Lodge (Motel) (Australia) deals
scrape_timestamp	When the run started (UTC).	2026-09-25T12:59:15Z
proxy_country	The country it was loaded from.	US
hotel_metadata	The full snapshot: review score, count and breakdown, photo count, neighbourhood, address, property type, description, every hotel facility, star rating, Booking.com id, coordinates, check-in time, sustainability badge.	{ "review_score": 7.5, "hotel_facilities": [ … ], … }

---

❓ FAQ

Which price should I compare on? nightly_total_price for a like-for-like comparison, lowest_bookable_rate for the price a guest sees first. Read tax_clarity_label before treating either as final.

Why is the price different from what I see on Booking.com? Usually one of four things: you see it from another country (proxy_country), with tax included (tax_clarity_label), as a Genius member (is_genius_rate), or in another currency.

Why are so many rows sold out? Because Booking.com shows those room types as unavailable for that night and party. On busy nights most room types of a small property are, and those rows are what the demand read is built on. Turn off includeSoldOut if you only want what is on sale.

What does rooms_left: null mean? Unknown, not plenty. Booking.com shows a count only when it wants to push a booking.

How often should I run it? Once a day, at the same hour. tieredCadence keeps long horizons affordable.

Can it scrape more than 10 hotels? Yes. Use searchUrl or searchText with maxProperties. Around 10 per run is the sweet spot for speed and reliability.

---

📝 What's new

👋 Used this Actor before September 2026?

This Actor had a large update in September 2026. Nothing you use was removed: every field, every input and your saved tasks keep working. What you will notice:

• New on every row: what changed since the last run (price_move_label, price_change, availability_change, days_since_price_change…), the cheapest bookable rate (lowest_bookable_rate), is_my_hotel and sort_key. 31 new fields in all, each in the field reference above.
• New outputs: Market by date (your rank against the others, night by night) and the daily digest by email or webhook.
• New inputs, all optional: myHotelUrl, trackPriceChanges, historyStoreName, digestEmail, digestWebhookUrl, childrenAges.
• Read differently now:
  • rooms_left is filled only when Booking.com prints "Only N left", so it is empty on most rows, and availability_status reads available unless such a count is shown.
  • nightly_total_price is the price as Booking.com displays it: tax_clarity_label says whether tax is added on top.
  • photo_count is empty when unknown, instead of 0.
  • los_prices holds the prices for booking 1, 2, 3… of a room at once.
  • With no startDate, the first night is today at the hotel.
• Your first run after the update shows ✦ New in the change columns: they compare each run with the previous one.
• The screens: Rate shopping still opens first, Rate shopping V2 is the daily screen, and the API views room_comparison and ai_room_picker still answer.

The last three releases

CHANGELOG.md

---

💬 Support

A question, a number that looks wrong, or an idea? Open an issue on the Issues tab. Every message is read.

Built by HotelSignals Lab.