Zillow Property Detail Scraper

This listing publishes the measured number for each field: how often it actually shows up on a real run, not how often it should.

What it does

Feed it Zillow listing URLs, street addresses, or Zillow IDs (zpids), on their own or chained from Zillow Search Scraper's dataset. Each input is one plain HTTP fetch with a Chrome TLS fingerprint, no browser process, returning the full property record: address, price, size, broker and agent detail, tax and estimate data where published, and the property's photos, all in one row. No second call for photos.

Which Zillow pages work

• Single home page (/homedetails/): works, full data, what most of the numbers below are measured against.
• Apartment building (/apartments/): works. One tested building returned 24 units, rents from $810 to $1,200, 1 to 2 bedrooms, the leasing office phone, and 24 photos.
• Builder community (/community/..._plid/): works. Of four live communities tested, three returned full rows: builder name, builder phone, price range, spec home and floor plan counts, and 11 to 21 photos. One had zero homes left and still returned the builder's name, phone, and photos. A separate community URL, the one the most-used Zillow detail Actor on the Store (about 1,000 monthly users) ships as its prefilled example, returns HTTP 404.
• Building page (/b/...): weak. Two tested came back nearly empty, no building name, no price, 0 to 1 units, 0 to 7 photos. Not worth running yet, and this listing will not tell you otherwise.
• /b/.../lat,long_ll/: refused on purpose. Zillow's robots.txt disallows this exact route, and at least one other Zillow Actor on the Store ships that same URL as its prefilled example anyway.

What comes back, and how often

Measured across 745 properties in Austin TX, Chicago IL, Denver CO, Greenville SC, Miami FL, and Phoenix AZ, in three runs, all chained from a search Actor dataset. Zero failed requests in all three.

Always present on all 745: street address, city, state, ZIP, Zillow ID, listing URL, price, home status, home type, currency, and photo count.

Shows up on	Field
99%	latitude and longitude
99%	description text
99%	bathrooms
99%	bedrooms
99%	at least one photo
99%	living area (square feet)
96%	MLS name
95%	MLS ID
88%	lot size
84%	listing agent phone
84%	broker name
83%	days on Zillow
82%	last sold price
82%	property tax rate
81%	listing agent name
79%	year built
74%	parking capacity
70%	date sold
36%	monthly HOA fee
28%	price change and its date
21%	rent Zestimate
19%	Zestimate
4%	county

Zillow publishes a Zestimate on about one property in five, so if your workflow depends on it, plan for four rows in five to be empty.

County shows up on only 4% of listing pages, worse than any other field here. If you need county on every row, plan to join it from another source.

Five fields, broker phone, broker office name, agent license number, stories, and tour view count, were cut from the output in build 0.1.19. Zillow carries the keys but never fills them, confirmed empty across all 745 properties tested. This listing counts 60 fields, not 65.

Photos

Across the 745 properties, this Actor returned 26,519 photo URLs, 36 per property on average, largest single property 60. 735 of 745, 99%, came back with at least one photo. 10 had none.

In the Austin batch, 50 requests were refused with an HTTP 403, retried automatically on a fresh IP, and all 150 properties still came back. This Actor bills per property returned, not per request, so a retry is never a separate charge. Every photo comes from that same fetch, no add-on charge.

When Expand building units is on, each unit row carries the parent building's photos, so a 24 unit building repeats the same 24 photo URLs across 24 rows.

How to use it

1. Give it URLs, addresses, or Zillow IDs. Paste a listing URL, type a street address, or supply a zpid. Chain it from Zillow Search Scraper's output dataset, or run it standalone.
2. Set Maximum items as your spend cap, the number of properties the run pulls before it stops.
3. Turn on Expand building units for an apartment building or builder community page, if you want a row per unit instead of one summary row.
4. Click Start, then read results from Storage or pull them through the API.

What you are charged for

You pay per property returned to your dataset. Price per 1,000 properties, by your Apify plan:

Your plan	Discount tier	Price per 1,000 properties
Free	No discount	$2.40
Starter	Bronze	$2.00
Scale	Silver	$1.60
Business	Gold	$1.50

Platform usage, compute, proxies, and storage, is included. No separate start fee, and no field is gated behind a paid add-on. This Actor bills per property returned, so a retry or a blocked request that gets auto-retried is never a separate line item.

The most-used Zillow detail Actor on the Store, at about 1,000 monthly users, charges $3.60, $3.00, $2.40, and $1.70 at those same tiers. This Actor is 33% cheaper at Free, Bronze, and Silver, and 12% cheaper at Gold.

A photo-only Actor on the Store charges $0.000727 per photo. At the 36 photos per property measured here, that is $0.0262 for photos alone, against $0.0020 total at Bronze here, photos plus the other 59 fields.

Input example

The realistic case: pull everything a search run already found, then top it up with a couple of specific listings, capped at a spend limit.

{
  "searchResultsDatasetId": "aBcD1234EfGh5678",
  "startUrls": [
    { "url": "https://www.zillow.com/homedetails/16210-S-14th-Dr-Phoenix-AZ-85045/8134236_zpid/" },
    { "url": "https://www.zillow.com/homedetails/5014-W-Frances-Pl-Austin-TX-78731/29342650_zpid/" }
  ],
  "maxItems": 500,
  "photoMaxWidth": "768",
  "expandBuildingUnits": false,
  "respectRobotsTxt": true,
  "proxyConfiguration": { "useApifyProxy": true }
}

searchResultsDatasetId pulls the zpid or URL off every row in a finished Zillow Search Scraper run and fetches the full detail page for each one. startUrls adds specific listings on top of that dataset, or stands alone if you skip the chained dataset entirely. maxItems caps the run at 500 properties, your spend limit, regardless of how many URLs are queued behind it.

Output example

A real row from the Phoenix run, unedited:

{
  "pageType": "property",
  "zpid": 8134236,
  "url": "https://www.zillow.com/homedetails/16210-S-14th-Dr-Phoenix-AZ-85045/8134236_zpid/",
  "homeStatus": "FOR_SALE",
  "homeType": "SINGLE_FAMILY",
  "price": 749000,
  "zestimate": 730300,
  "rentZestimate": 4137,
  "lastSoldPrice": 410000,
  "dateSold": "2017-11-14",
  "propertyTaxRate": 0.41,
  "monthlyHoaFee": 51,
  "totalMonthlyCost": 4514,
  "bedrooms": 5,
  "bathrooms": 3,
  "livingArea": 2839,
  "lotSize": 8594,
  "yearBuilt": 1993,
  "parkingCapacity": 3,
  "streetAddress": "16210 S 14TH Drive",
  "city": "Phoenix",
  "state": "AZ",
  "zipcode": "85045",
  "latitude": 33.29913,
  "longitude": -112.09028,
  "daysOnZillow": 4,
  "mlsId": "7071822",
  "mlsName": "ARMLS",
  "pageViewCount": 299,
  "favoriteCount": 36,
  "agentName": "Kevin Houston",
  "agentPhone": "602-370-8106",
  "brokerName": "West USA Realty",
  "photoCount": 43,
  "photos": [
    "https://photos.zillowstatic.com/fp/8ff578e47e35892a7842ea9fc38ec734-cc_ft_1536.jpg",
    "https://photos.zillowstatic.com/fp/0d9f91d5e83ea8bc9b39def2c1d42a00-cc_ft_1536.jpg",
    "... 40 more"
  ],
  "hasVirtualTour": false,
  "description": "Gorgeous 5 bed, 3 bath in Club West located on a very quiet, low traffic street! ..."
}

One row per property, with photos as an array on that same row. A building or community page returns a differently shaped row for the same run: priceMin, priceMax, and unitCount for an apartment building, or builderName and homeCount for a community, in place of the single-home fields above.

Integrations

Every public Actor on Apify carries the platform's integrations, so a run can push property rows wherever your team already works.

• n8n, Make, Zapier, and Pipedream for no-code automation.
• HubSpot and Airtable to land properties as CRM records or rows.
• Slack, Telegram, and Gmail to push properties to people rather than to storage.
• Google Drive, Snowflake, Keboola, and Airbyte for exports and warehouse loads.
• Webhooks and the GitHub integration for anything custom.
• LangChain, the OpenAI Agents SDK, the Vercel AI SDK, and Google ADK to hand property records to an LLM app.

Two combinations worth setting up on day one:

A watchlist that only reports changes. Switch on Only new or changed properties, give it a memory store name, schedule the run, and point a webhook or Slack step at the finished dataset. A property that has not changed since the last run is skipped and never charged.

Search feeding detail. Run Zillow Search Scraper cheaply over an area, then pass its dataset id into this Actor's Search results dataset field to pull the full record, including photos, for everything the search found.

FAQ

Is it legal to scrape Zillow?

The data on a Zillow property page is public, reachable with no login, and the only personal contact detail returned is what a listing agent already publishes as business contact information. You are responsible for how you use what you collect. Read Apify's guide on the legality of web scraping.

Does this Actor respect robots.txt?

Yes, and it is on by default here, unlike the sibling search Actor, where a filtered search has no compliant route and the switch defaults off. Zillow's robots.txt permits /homedetails/, /apartments/, and /community/. It disallows /b/*/*_ll/, and those URLs are skipped, with a line in the run log explaining why.

Why is the Zestimate missing on most rows?

Zillow only publishes a Zestimate on about one property in five, measured at 19% across 745 properties tested. This Actor passes the value through exactly as Zillow gives it and never estimates one of its own.

Can I get one row per photo?

No, and you do not need one. Photos come back as an array on the property row, and Apify's CSV and Excel export flattens that array into numbered columns, photos/0, photos/1, and so on. Actors that bill per photo charge many times more for the same single page fetch.

What happens to a URL that no longer exists?

A dead listing returns HTTP 404, is logged, and is not charged. An address that cannot be matched to a listing is reported in the run log and skipped, rather than failing the whole run.

Can I use this with the Apify API or an MCP server?

Yes, both. The Actor runs under limited permissions, which is why chaining a search run uses a dataset picker that declares read access, rather than a pasted dataset id.

Support

Missing a field, or want a different output shape? Open an issue on this Actor's Issues tab. Field requests are welcome. If Zillow publishes it, it can usually be added.

To find properties in the first place, see Zillow Search Scraper. To work from the agent side instead, listing the estate agents in an area and everything each one has on the market, see Zillow Agent Scraper.