Zillow ZIP Market Scraper avatar

Zillow ZIP Market Scraper

Pricing

from $1.76 / 1,000 listing scrapeds

Go to Apify Store
Zillow ZIP Market Scraper

Zillow ZIP Market Scraper

Scrape Zillow search results by ZIP, city or URL for forSale, forRent and recentlySold. Reads the JSON search API (3.2 KB per listing), extracts full hdpData.homeInfo with zestimate, rent Zestimate, tax assessed value and days on Zillow, and auto-splits price bands above 800 listings.

Pricing

from $1.76 / 1,000 listing scrapeds

Rating

0.0

(0)

Developer

Mohit Tripathi

Mohit Tripathi

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

10 hours ago

Last modified

Categories

Features

Limited permissions

Share

Scrape Zillow search results by ZIP, city, lat,lng or Zillow URL for forSale, forRent and recentlySold. Reads Zillow's JSON search API directly — no browser — and returns one row per listing with the full hdpData.homeInfo block.

  • Pay per event: listing-scraped $2.00 / 1,000 rows + apify-actor-start $0.05 / 1,000 starts.
  • 3.19 KB per listing on the JSON API vs 15.62 KB/listing on the HTML __NEXT_DATA__ route (and the HTML route is 403 for datacenter and residential clients).

Input

FieldTypeDefaultNotes
locationsarray["10001"]ZIP, city/state, lat,lng, or Zillow search URL.
statusstringforSaleforSale, forRent, recentlySold.
priceBandsarrayautoOptional {"min":0,"max":250000} bands; auto-generated above 800 listings.
maxListingsinteger500Hard stop; billing stops here.
maxPagesPerBandinteger20Zillow caps pagination at ~20 pages.
requestTimeoutSecsinteger30Per request.

Prefill (10001, for sale, 1 page) runs in ~5.5 s.

Output

{
"zpid": "72515759",
"status": "Condo for sale",
"price": 1999000,
"price_text": "$1,999,000",
"currency": "USD",
"price_period": "total",
"address": "133 W 17th St APT 3A, New York, NY 10011",
"city": "New York",
"state": "NY",
"zip": "10011",
"latitude": 40.740005,
"longitude": -73.996635,
"home_type": "CONDO",
"bedrooms": 2,
"bathrooms": 2,
"living_area": 1400,
"lot_size": null,
"lot_size_unit": null,
"year_built": null,
"days_on_zillow": 7,
"zestimate": 1739100,
"rent_zestimate": 8431,
"tax_assessed_value": null,
"price_change": null,
"price_history": null,
"is_building": false,
"is_zillow_owned": false,
"image_url": "https://photos.zillowstatic.com/fp/8a1f015a912a338aff11c00d5465e7ee-p_e.jpg",
"url": "https://www.zillow.com/homedetails/133-W-17th-St-APT-3A-New-York-NY-10011/72515759_zpid/",
"price_band": null,
"source": "zillow.com",
"scraped_at": "2026-11-20T09:40:12.000000+00:00"
}

Rentals use price_period: "monthly" and price = monthly rent. Apartment-building cards (is_building: true) have no hdpData; their price is minBaseRent.

Cost

Volumelisting-scrapedActor startsTotal
1,000 listings$2.00$0.05$2.05
10,000 listings$20.00$0.05$20.05

No charge for blocked pages, empty pages, duplicates, or rows without zpid plus price/address.

Source notes (verified)

  • HTML search pages (/homes/10001_rb/) return 403 for both the datacenter IP and residential proxies — __NEXT_DATA__ is not a usable route.
  • PUT https://www.zillow.com/async-create-search-page-state returns JSON with cat1.searchResults.listResults and categoryTotals.cat1.totalResultCount: 41 listings, 133,898 bytes → 3.19 KB/listing. That is 80% smaller than the 15.62 KB/listing HTML figure.
  • searchQueryState needs usersSearchTerm plus mapBounds plus isMapVisible: true; a term alone returns zero results, and Zillow resolves the region from the map box (not the term).
  • Status filters: the full isForSaleByAgent / isForSaleByOwner / isNewConstruction / isComingSoon / isForSaleForeclosure / isForRent / isRecentlySold block must be sent; a single isForRent flag returns nothing.
  • Price filter: filterState.price = {"min": ..., "max": ...} works (verified 0–250k → 8 results, 1M–1.5M → 476 results).
  • Pagination: pagination.currentPage; 41 per page, ~20 pages max.
  • ZIPs are geocoded to a centroid with zippopotam.us, then searched in a ±0.008° lat / ±0.010° lon box and filtered to the exact ZIP. Cities use Nominatim's bounding box.
  • Direct requests are blocked; every request goes through Apify Residential proxy with 5 attempts, 3–15 s backoff. A blocked page is skipped and never billed.

Price-band splitting

When totalResultCount > 800 (and maxListings > 200):

  1. Probe the location, then query bands 0–250k, 250–500k, 500–750k, 750k–1M, 1M–1.5M, 1.5M–2M, 2M+ (rentals: 0–1k, 1–2k, 2–3k, 3–5k, 5k+).
  2. Any band still above 800 is split in half again, up to depth 3.
  3. Results merge and dedupe by zpid.

Verified on Austin, TX: 7,514 for-sale listings → bands split (0–250k = 827 → 0–125k / 125k–250k, 250–500k = 3,757 → 250–375k / 375–500k, …), 250 unique listings returned with no duplicates.

MCP

{
"actor": "structlabs/zillow-zip-market-scraper",
"input": {"locations": ["10001", "Austin, TX"], "status": "forSale", "maxListings": 250}
}

Local verification

  • 20-input gate with ACTOR_TEST_PAY_PER_EVENT=true apify run: 18/20 returned rows, 20/20 clean (zero exceptions, zero duplicates, correct ZIP on every ZIP run). The two zero-row runs were ZIP 30301, a PO-box-only ZIP with no Zillow inventory; the same cases with 30305 returned 41 and 45 rows.
  • Median 10.9 s per run; the 250-listing banded Austin run took 89 s.
  • hdpData.homeInfo fields (zestimate, rentZestimate, taxAssessedValue, daysOnZillow) present wherever Zillow publishes them.

Pricing source

Canonical store-discount ladder (FREE/BRONZE/SILVER/GOLD/PLATINUM/DIAMOND) lives in pricing.json. Apify actor.json cannot carry pricing and apify push does not change account pricing; after any manual API change, the newest pricingInfos record must match pricing.json.

n8n Workflow & MCP Quickstart

Run this Actor from an AI agent over MCP, or import the ready-made n8n workflow: