Booking Hotel Finder & Scraper: AI best-value shortlist on the Apify Store →
Tutorial article for the Apify blog guest program, dev.to, Medium and Hashnode. Drafted with the listing; refreshed with the first real run by /actor-marketing; the owner publishes it and records the URL in brief.json → marketing.links.
A travel agent with a client flying to Lisbon for two nights does the same thing every time: open Booking.com, set the dates, sort by review score, open fifteen tabs, divide each stay price by the nights, scroll every listing for "free cancellation", check whether the pool is real or "nearby", and read the fine print for a city tax or a cash-only desk. A corporate-travel planner does it again per conference, a deal hunter every morning. It is an hour per destination, and a scheduled Booking scraper does not help: it returns 100–1,000 raw rows in search order — sold-out and over-budget properties included, billed per row — and tomorrow it returns the same hotels again.
Booking Hotel Finder turns that hour into one run: it runs the Booking Scraper (voyager/booking-scraper) on your Apify account, drops junk for free, ranks what is left by value per night and — with the AI tier — checks each hotel against a plain-language traveller brief with quotes from the listing.
1. Try the free demo. Open the Actor's Store page and click Try it. The form is pre-filled with six sample Lisbon hotels (sourceMode: list) and runs in seconds, nothing charged. Two of the six never reach the result: a hostel rated 7.1 (below minRating 7.5) and a riverside loft with no room left for the dates (sold out). Both are free filters — you would never pay for them.
2. Read the Shortlist. The dataset opens on the Shortlist view (docs/img/shortlist.svg): rank, hotel, price per night, rating, value score, value tier, free cancellation, distance to your point and a why-this-hotel line. In the demo, Hotel Miradouro da Graça ranks first: $124/night, the cheapest in the search, rated 8.6 from 2,410 reviews, 0.79 km from Rossio. The Client sheet view (docs/img/client-sheet.svg) shows the pros and cons from Booking's review summary and the fine print: "A damage deposit of EUR 100 is required on arrival and refunded at check-out."
3. Search your own stay. Switch Where do the hotels come from? to Search Booking.com live, type a destination, checkIn and checkOut, and optionally:
maxPricePerNight — your nightly budget, sent to Booking's price filter and checked again;mustHaveAmenities — e.g. ["pool", "parking"], matched against Booking's structured facility list with synonyms;nearPoint + maxDistanceKm — the venue or old town as latitude,longitude;travellerBrief — "Quiet hotel for a couple, walkable to the old town, with a pool and free cancellation".Leave enableAi on: your first 10 hotels are free, so the first run shows the AI tier at no cost.
4. Export or wire it. Export the view as CSV / Excel, connect Google Sheets from the Integrations tab, or set webhookUrl to receive each hotel with a Slack-ready line: ↓12 % Hotel X, $143/night, 8.7 Excellent.
On the demo's Baixa Garden Suites the rule tier already gives value score 68, $149/night, rating 8.9 from 640 reviews and a keyword brief match ("Rooftop pool" for pool). The AI tier reads the brief "Quiet hotel for a couple, walkable to the old town, with a pool and free cancellation" and returns briefMatchScore 75 with evidence copied from the listing — "300 metres from Rossio Square." for walkable to the old town, "shared rooftop pool" for pool — and marks quiet hotel for a couple as not proven. It writes a why-this-hotel line ("Only $149/night, 0.14 km from Rossio Square, rooftop pool and free cancellation – a central stay for couples.") and fine-print flags such as "late check-in carries an extra charge of EUR 30." and "This property only accepts cash for the city tax." Every quote must be a substring of the listing or it is dropped; a number that is not in the data sends the line back to the rule version. (Local demo runs, 2026-09-26: 2.8 s with AI off, 6.9 s with AI on.)
100 properties scraped by voyager/booking-scraper cost $0.50 on your account ($0.30 with santamaria-automations/booking-com-scraper), capped by maxDiscoveryChargeUsd. Say 25 pass your filters and the value bar: on the first run 10 of them are free (15 × $0.004 = $0.06). On later runs they cost 25 × $0.004 = $0.10 with AI off — $0.60 all in, $0.024 per shortlisted hotel — or 25 × $0.01 plus ≈ $0.003 of tokens each with AI on: $0.83 all in, $0.033 per hotel. Sold-out, over-budget, low-rated, duplicate and unchanged hotels are never charged, and there is no start fee.
For price-drop alerts, save the input (fixed or relative dates) and add a daily schedule (Actions → Schedule) with your address in notifyEmail. onlyNewOrPriceDrop (on by default) remembers each delivered stay and its price, so the next run delivers — and charges — only new hotels or drops of minPriceDropPct (5 %) or more, with previousPrice and priceChangePct. From code:
from apify_client import ApifyClient
client = ApifyClient("<YOUR_API_TOKEN>")
run = client.actor("rich_minds/booking-hotel-finder-ai").call(run_input={
"sourceMode": "actor", "destination": "Lisbon", "checkIn": "2026-10-26", "checkOut": "2026-10-28",
"travellerBrief": "Quiet couple trip with a pool", "maxDiscoveryChargeUsd": 0.5,
}, timeout_secs=1800)
for hotel in client.dataset(run["defaultDatasetId"]).iterate_items():
print(hotel["rank"], hotel["name"], hotel["pricePerNight"], hotel["whyThisHotel"])A daily watch of 100 properties with about five new or cheaper hotels a day is ≈ $0.57 a day — about $17 a month, against rate-shopper seats such as Lighthouse Rate Insight from about $250 a month per property (indicative list price; a rate shopper also does 365-day room-type calendars, which this Actor does not).
Open Booking Hotel Finder on the Apify Store, click Try it and run the pre-filled demo — free, in about three seconds. Then type your own destination and dates; your first 10 qualified hotels are free. If you use it for clients, a short review on the Store page helps other travel planners find it.
Disclosure: written by the developer of the Actor.
Booking Hotel Finder & Scraper: AI best-value shortlist on the Apify Store →