Realitica Property Scraper & Listing Monitor

Use this Realitica property scraper to collect listings across Balkan markets as structured rows, then reuse the same input as a Balkan property monitor for new listings, updates, price movements, and safely proven removals. The Actor also produces free coverage evidence, market summaries, duplicate-agency clusters, and explainable comparable-property rows.

Start with familiar location fields or paste exact Realitica URLs. Basic runs stay lightweight; richer detail pages are opened only when you request them or activate a detail-dependent filter.

🚀 Start in 60 seconds

1. Open Input and choose a country, area, listing mode, and property type.
2. Keep Output mode on Current snapshot for the first exploratory run.
3. Set Maximum listing rows to a small number such as 50.
4. Click Start, then open the Property listings dataset view.

{
  "countries": ["Crna Gora"],
  "areas": ["Tivat"],
  "listingModes": ["sale"],
  "propertyTypes": ["Apartment"],
  "maxListings": 50,
  "stateMode": "current"
}

If you already have an exact Realitica results page or listing page, use startUrls instead:

{
  "startUrls": [
    { "url": "https://www.realitica.com/hr/listing/3069833" }
  ],
  "maxListings": 50
}

🎯 Choose the right workflow

What you need	Recommended input	What happens
Browse a market	Country, area, mode, and property type	The Actor builds bounded Realitica searches
Reuse an exact source scope	startUrls	Up to three supplied URLs replace the generated searches
Export all current listings	stateMode: "current"	Every useful listing row is returned
Monitor only useful changes	stateMode: "changes"	First run seeds; later comparable runs emit changes
Bedrooms, parking, sea distance, seller, or phone	Detail controls	Detail pages open automatically and successful enriched rows use the detail price
Compare one property with the fetched market	referenceListing	Free comparable rows explain each similarity score

When startUrls contains at least one URL, it takes precedence over Countries, Areas, Listing modes, Property types, source price, posting-window, and sort fields. Local row filters still apply after the supplied pages are fetched.

Generated searches are limited to three country × listing-mode targets per run. For example, one country with three modes is valid; two countries with two modes would create four targets and is rejected before source work.

📦 What you get

The default dataset intentionally contains several useful record types. Use recordType to select the right rows:

recordType	Purpose	Recommended view
listing	Structured current listings or change events	Property listings
market_summary	Count, price, and price-per-m² statistics	Analysis
duplicate_cluster	Likely cross-agency duplicates with matching reasons	Analysis
comparable	Reference-property matches with explainable scores	Analysis
coverage	Target results, limits, warnings, field coverage, and billing summary	Coverage

Listing rows can include location, property type, mode, price, currency, surface, price per m², rooms, seller, source dates, images, state fields, and the canonical Realitica URL. Detail rows may additionally contain bedrooms, bathrooms, parking, distance from sea, description, phone, and a list of fields actually delivered.

Core export fields are listingId, canonicalUrl, title, country, city, propertyType, listingMode, price, currency, surfaceM2, postedDate, postedDatePrecision, updatedDate, updatedDatePrecision, firstSeenAt, lastSeenAt, changeStatus, coverageState, coverageStatus, and collectedAt. Source-dependent fields remain null when the fetched page does not expose them.

{
  "recordType": "listing",
  "listingId": "example-id",
  "title": "Apartment in Tivat",
  "listingMode": "sale",
  "price": 180000,
  "currency": "EUR",
  "surfaceM2": 60,
  "pricePerM2": 3000,
  "changeStatus": "CURRENT",
  "coverageState": "COMPLETE"
}

For clean exports, filter recordType == "listing"; use the dedicated Analysis and Coverage views for the other rows. The OUTPUT key-value-store record provides a buyer-facing delivery summary, coverage, state counts, resource totals, and charge totals.

🎛️ Input guide

Location search. Countries, Areas, Listing modes, and Property types build the source queries. Source price, date, and sort controls narrow generated searches. Labels should match Realitica's current catalog.

Exact URLs. startUrls is an alternative source selector, not an additional filter. Once supplied, it replaces the generated location searches. Result limits and local filters still apply.

Local filters. Rooms, surface, price per m², currency, keywords, and sublocations filter normalized rows. A completed source request can therefore produce FILTERED coverage when valid source rows do not pass the selected local rules.

Paid detail activation. includeDetails, includeContact, parking: true, bedroom bounds, maximum sea distance, or a seller query opens detail pages automatically. A successfully enriched row receives one detailed_listing_result event instead of a basic event. Failed or unavailable detail can fall back to a basic row when delivery capacity remains.

Comparables. referenceListing scores only the listings already fetched in the run. It does not launch a broader premium search, and comparable evidence rows are free.

Limits. maxListings is the global listing-row ceiling. maxPagesPerSearch limits each source search. Reaching either while more data may remain produces CAPPED coverage rather than a false complete result.

Repeat runs & monitoring

Choose Changes since the previous comparable run, then repeat the same row-affecting input sequentially.

• The first Changes run emits observed listings as SEEDED and creates the baseline.
• Later comparable runs emit NEW listings and meaningful UPDATED rows, including price-change fields.
• Enable includeUnchanged only for audit exports; unchanged rows are free.
• A missing listing becomes ENDED with availabilityStatus: "REMOVED" only after two comparable complete absences.
• Capped, partial, blocked, failed, or filtered snapshots do not prove removal.

The Actor derives comparison scope from the source, filters, and detail options. Changing one of those settings creates a separate scope. Avoid overlapping runs for the same monitor.

💰 Pricing

This Actor uses pay per event. The configured schedule charges one start only after a usable verified source result or positively confirmed exhausted-empty result, plus one mutually exclusive event for each useful listing row.

run charge = $0.005 verified actor start
           + delivered basic rows × your plan's listing_result price
           + delivered detailed rows × your plan's detailed_listing_result price

Event	Free tier	Bronze	Silver	Gold
Validated Actor start (actor_start)	$0.005	$0.005	$0.005	$0.005
Basic listing or useful change (listing_result)	$0.00080	$0.00072	$0.00064	$0.00056
Detailed listing or useful change (detailed_listing_result)	$0.00160	$0.00144	$0.00128	$0.00112

Free-tier examples:

• 25 basic Current rows: $0.005 + 25 × $0.00080 = $0.025
• 25 successfully detailed Current rows: $0.005 + 25 × $0.00160 = $0.045
• A successful verified Changes run with no paid change rows: $0.005

The detailed event replaces the basic event for that row; it is not added on top. Blocked, failed, and unverified source attempts do not charge the start. Unchanged listings, coverage, summaries, clusters, comparables, and warnings are free. If requested detail is not successfully delivered, an otherwise useful row can fall back to the basic event.

✅ Coverage you can trust

Coverage evidence distinguishes a real source result from a stopped, filtered, or unverified run:

• COMPLETE — the configured source window completed without a known cap.
• EMPTY_CONFIRMED — the source window completed and independently supports zero matching rows.
• CAPPED — a page, row, request, runtime, or billing ceiling stopped broader work.
• PARTIAL — useful data was delivered, but part of the requested work failed or remained unresolved.
• FILTERED — source rows were found, but local filters removed them from listing delivery.
• BLOCKED — the source denied reliable access for that target.
• FAILED — no trustworthy result or verified empty result was obtained.

Coverage rows remain in the mixed dataset even when no paid listing row is emitted. targetResults, warnings, raw and unique counts, duplicates removed, and field-coverage statistics explain what happened.

🔌 API and automation

Run with the Apify API by replacing ACTOR_ID and APIFY_TOKEN:

curl -X POST \
  "https://api.apify.com/v2/acts/ACTOR_ID/runs?token=APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"countries":["Crna Gora"],"areas":["Tivat"],"listingModes":["sale"],"maxListings":50}'

Export dataset rows as JSON, CSV, Excel, XML, RSS, or JSONL through Apify dataset endpoints. For monitoring, first save the exact Changes input as an Apify Task, run it once to seed state, and only then add a non-overlapping schedule.

❓ FAQ and support

Why were my country and area settings ignored?
Check startUrls. Any supplied Realitica URL takes precedence over the generated location search.

Why did a basic-looking filter increase the row price?
Bedrooms, parking, sea distance, seller, and phone are detail-page fields. Requesting one automatically activates detail work. Successful enriched rows use the detailed price instead of the basic price.

Why do some columns look empty?
The dataset contains listings, summaries, clusters, comparables, and coverage. Filter on recordType, or switch to the matching dataset view.

Why did the first Changes run return SEEDED rows?
It created the baseline. Keep the same source, filters, and detail options, then run sequentially to receive later changes.

Why was a missing listing not marked REMOVED?
Removal needs two comparable complete absences. A capped, partial, blocked, failed, or filtered run cannot safely prove disappearance.

For support, share the Apify run ID, the input with sensitive values removed, and the Coverage row or buyer-facing OUTPUT warning.