Halooglasi.com Property Scraper

Scrape real estate listings from Halooglasi.com — Serbia's #1 property portal — into a clean, structured Apify dataset.

This actor extracts apartments, houses, land, commercial space, garages and rooms (sale or rent) by location, transaction type, EUR price, m², rooms and owner type. Each record includes price (EUR), surface area, full Serbian address (city → municipality → district → micro-location), advertiser, image gallery and listing labels.

---

Features

• 🇷🇸 Serbia-wide coverage — every Halooglasi location slug is supported (Beograd municipalities, Novi Sad, Niš, Kragujevac, etc.).
• 💰 Native EUR pricing with auto-computed pricePerSqm.
• 🏘️ Four-level location chain — city, municipality (opština), district, and micro-location parsed separately.
• 🛏️ Serbian room-count handling in 0.5 increments (garsonjera, jednosoban, dvosoban, troiposoban, …) returned as both numeric value and Serbian label.
• 🏢 Advertiser data — agency name, profile URL, agency ID and owner type (Agencija / Vlasnik / Investitor).
• 🖼️ Image gallery per listing (logo banners auto-filtered).
• 🏷️ Card tier labels (Premium, Top, VIP, …).
• 🛡️ Cloudflare-resilient — direct HTTP fetch via residential proxy with automatic retry on 403/429/503.
• 🔁 De-duplication & loop detection across paginated pages.

---

Architecture

Halooglasi is a Sitecore + ASP.NET site. Each listing page server-side embeds an inline JS variable carrying the full ad payload:

window.QuidditaEnvironment.serverListData = {
  Ads: [...],          // up to 20 ads per page
  TotalCount: 9198,
  TotalPages: 460,
  ...
};

The actor:

1. Regex-extracts the serverListData JSON blob from the response HTML.
2. Walks Ads[] — each ad gives Id, Title, RelativeUrl, AdvertiserId, and ListHTML.
3. HTML-entity-decodes ListHTML (the visible product card markup) and parses with cheerio to extract price (from <span data-value>), area, rooms, floor, location chain, images and advertiser.
4. De-duplicates by adId across pages and detects pagination loops.

If the JSON envelope is missing (rare), a cheerio-based HTML fallback parses .product-item cards directly.

No browser, no GraphQL, no Cloudflare-bypass tricks — just got-scraping + cheerio over residential IPs.

---

Input

Field	Type	Default	Notes
locationSlugs	string[]	["beograd"]	One slug per task. Examples: beograd, novi-sad, nis, kragujevac, beograd-novi-beograd, beograd-vracar, beograd-savski-venac-beograd-na-vodi, beograd-zvezdara, beograd-vozdovac, beograd-cukarica, beograd-zemun, beograd-palilula, beograd-stari-grad. Find more by browsing halooglasi.com and copying the path between /nekretnine/{txn}/ and ? from the URL.
transaction	sale | rent	sale	Maps to prodaja / izdavanje.
propertyType	enum	apartment	apartment (stan) / house (kuća) / land (zemljište) / commercial (poslovni prostor) / garage (garaža) / room (soba).
priceMin / priceMax	int (EUR)	0	0 = no bound.
areaMin / areaMax	int (m²)	0	0 = no bound.
roomsMin / roomsMax	num	0	Halooglasi uses 0.5 increments (garsonjera = 0.5, dvosoban = 2, 5+ = 5). 0 = no bound.
ownerType	enum	all	all / agency / owner / investor.
sort	enum	latest	latest / price_asc / price_desc / area_asc / area_desc.
maxListings	int	200	Total cap across all tasks. 0 = unlimited.
maxPagesPerTask	int	10	Pagination depth per location (≈ 20 listings per page). Narrow with price/area for deeper inventory.
requestDelay	int (ms)	1000	Inter-request delay. ≥ 800 ms recommended.
maxRetries	int	3	Per-request retries on 403 / 429 / 503 (rotates proxy IP).
proxyConfiguration	proxy	RESIDENTIAL + RS	Residential proxy strongly recommended — Halooglasi sits behind Cloudflare and datacenter IPs frequently get blocked. Country RS is prefilled.

Example input

{
  "locationSlugs": ["beograd-novi-beograd", "beograd-vracar"],
  "transaction": "sale",
  "propertyType": "apartment",
  "priceMin": 80000,
  "priceMax": 250000,
  "areaMin": 40,
  "areaMax": 80,
  "roomsMin": 1.5,
  "roomsMax": 3,
  "ownerType": "agency",
  "sort": "latest",
  "maxListings": 500,
  "maxPagesPerTask": 25,
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"],
    "apifyProxyCountry": "RS"
  }
}

---

URL pattern (decoded)

https://www.halooglasi.com/nekretnine/{prodaja|izdavanje}-{stanova|kuca|zemljista|...}/{location-slug}
  ?cena_d_from=<int>
  &cena_d_to=<int>
  &cena_d_unit=4                              ← EUR (RSD = different code)
  &kvadratura_d_from=<int>
  &kvadratura_d_to=<int>
  &kvadratura_d_unit=1                        ← m²
  &broj_soba_order_i_from=<num>
  &broj_soba_order_i_to=<num>
  &oglasivac_nekretnine_s=Agencija|Vlasnik|Investitor
  &order=1|3|4|5|6                            ← latest / price asc / price desc / area asc / area desc
  &page=<N>

---

Output

One Apify dataset record per listing:

Field	Type	Description
adId	string	Halooglasi 13-digit listing ID
detailUrl	string	Full listing URL
title	string	Listing title
shortDescription	string	Card-level description snippet
transactionType	string	sale / rent
propertyType	string	apartment / house / land / …
advertType	string	Raw Agencija / Vlasnik / Investitor
price	number	Total price in EUR
priceCurrency	string	EUR
pricePerSqm	number	Price per m² (read from card or computed price / areaSqm)
areaSqm	number	Usable surface area (kvadratura)
terrainAreaSqm	number	Plot/terrain area (houses, land)
rooms	number	0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5
roomsLabel	string	garsonjera / jednosoban / dvosoban / troiposoban / …
floor	string	Floor number/label (III, PR, VPR, …)
totalFloors	string	Total floors in building
buildingType	string	Stara gradnja / Novogradnja / …
heatingType	string	CG / EG / TA / podno / …
yearBuilt	int	Construction year
country	string	Srbija
city	string	City
municipality	string	Opština
district	string	District
microLocation	string	Mikrolokacija (neighborhood)
street	string	Ulica
fullAddress	string	Composed street, microLocation, district, municipality, city, Srbija
latitude / longitude	number	List endpoint does not expose GPS — always null (would need per-ad detail fetch).
mainImageUrl	string	Cover photo
imageUrls	string[]	All listing photos (agency logos filtered)
imageCount	int	Number of listing images
advertiserId	string	Halooglasi advertiser ID
advertiserName	string	Advertiser display name
advertiserUrl	string	Advertiser profile page
advertiserType	string	Agencija / Vlasnik / Investitor
isRegistered hasElevator hasParking hasGarage hasTerrace hasBalcony isFurnished	boolean	Detail-only on Halooglasi list endpoint — typically null (would need per-ad detail fetch).
isExclusive	boolean	True when card carries the Ekskluziva tier tag
datePosted	string	Most recent post date (when available)
dateValidFrom	string	Listing valid-from date (when available)
labels	string[]	Tier tags (Premium, Top, VIP, …)
searchTransaction searchPropertyType searchLocation searchUrl	string	Echoes of the input search parameters
scrapedAt	string	ISO-8601 scrape timestamp

Two dataset views are pre-configured:

• Overview — compact card-style table (ID, type, price, area, rooms, location, advertiser).
• Full Detail — every field above.

Sample record

{
  "adId": "5425647088741",
  "detailUrl": "https://www.halooglasi.com/nekretnine/prodaja-stanova/fontana-svetao-za-adaptaciju/5425647088741?kid=4",
  "title": "Fontana, svetao, za adaptaciju",
  "shortDescription": "Na prodaju izuzetno svetao stan na jednoj od najtraženijih lokacija – Fontana...",
  "transactionType": "sale",
  "propertyType": "apartment",
  "advertType": "Agencija",
  "price": 220000,
  "priceCurrency": "EUR",
  "pricePerSqm": 3548,
  "areaSqm": 62,
  "rooms": 2,
  "roomsLabel": "dvosoban",
  "floor": "IX",
  "totalFloors": "9",
  "country": "Srbija",
  "city": "Beograd",
  "municipality": "Opština Novi Beograd",
  "district": "Fontana",
  "microLocation": "Pariske komune",
  "fullAddress": "Pariske komune, Fontana, Opština Novi Beograd, Beograd, Srbija",
  "mainImageUrl": "https://img.halooglasi.com/slike/oglasi/Thumbs/260429/m/fontana-svetao-za-adaptaciju-...jpg",
  "imageCount": 1,
  "advertiserId": "5587725",
  "advertiserName": "BELIGRAD",
  "advertiserUrl": "https://www.halooglasi.com/oglasi/BELIGRAD",
  "advertiserType": "Agencija",
  "isExclusive": false,
  "labels": ["Premium"],
  "searchLocation": "beograd",
  "scrapedAt": "2026-05-05T14:22:17.392Z"
}

---

Important notes

• Cloudflare protection — Halooglasi explicitly blocks several AI crawlers in robots.txt and serves 403 from datacenter IPs. Use Apify Residential Proxy with country RS for reliable runs (prefilled in input). The actor automatically retries 403 / 429 / 503 responses with a fresh proxy session.
• Currency is always EUR (cena_d_unit=4). Halooglasi does internally support RSD but its UI defaults to EUR.
• Room counts are returned both as numbers (0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5) and as Serbian labels (garsonjera, jednosoban, dvosoban, troiposoban, …) for convenience.
• Coordinates and feature flags (lift, parking, garage, terrace, balcony, furnished, registered) are populated by Halooglasi's detail endpoint, not the list endpoint this scraper uses. They are kept in the schema (always null here) so downstream consumers can merge in detail-page enrichment if needed.
• Pagination cap — Halooglasi displays up to ~9,000 results per location; if you need a deep historical sweep, narrow with priceMin / priceMax / areaMin / areaMax ranges and run multiple tasks.
• Throughput — ~200 listings / ~70 seconds with residential proxy, including Cloudflare retries.

---

Common location slugs

Beograd municipalities — beograd-novi-beograd, beograd-vracar, beograd-savski-venac-beograd-na-vodi, beograd-zvezdara, beograd-vozdovac, beograd-cukarica, beograd-zemun, beograd-palilula, beograd-stari-grad, beograd-rakovica, beograd-mladenovac, beograd-grocka, beograd-obrenovac, beograd-surcin, beograd-lazarevac, beograd-barajevo, beograd-sopot, beograd-zvezdara-mirijevo

Other major cities — novi-sad, nis, kragujevac, subotica, pancevo, cacak, kraljevo, leskovac, zrenjanin, smederevo, valjevo, vranje, uzice, sabac, sombor, pozarevac, zajecar

To find more, browse a result page on halooglasi.com and copy the slug from the URL between /nekretnine/{txn-type}/ and ?.

---

Troubleshooting

Symptom	Likely cause	Fix
Repeated 403 warnings, 0 results	Datacenter proxy or no proxy	Switch to Residential with country RS
Two empty pages — stopping early	Filters too narrow for that location	Loosen price/area/rooms or split into multiple tasks
Pagination loop detected	Halooglasi caps result set ~9k per query	Narrow with price/area ranges to slice the inventory
floor / totalFloors null on some rows	Listing didn't include a floor feature	Expected — not every Halooglasi ad provides this field
latitude / longitude always null	List endpoint doesn't expose coordinates	Expected — would require per-ad detail fetch

---

License

Apache-2.0