Zillow Property Details Scraper

Zillow Property Scraper — Price, Zestimate, Tax History and Photos

Zillow Property Scraper pulls the full public record for any Zillow property — price, Zestimate, bedrooms and bathrooms, tax and price history, schools, nearby homes and photos — from a URL, a bare ZPID, or a plain street address. Every run returns clean, structured JSON on the same fixed schema every time, so field names never shift between runs. Paste one address or hundreds of URLs and ZPIDs in a single input list. Start a run on Apify and watch the first property land in your dataset.

🏠 What is Zillow Property Scraper?

Zillow Property Scraper is an Apify Actor that fetches a Zillow property's detail page plus its companion GraphQL modules and normalizes everything into one JSON record — price and facts, Zestimate, tax and price history, schools, nearby homes, photos, and listing information. No Zillow account or login is required: every field returned is exactly what a logged-out visitor can see on the public listing page. It's built for real estate developers, market analysts, and AI engineers who need structured property data without maintaining their own scraping pipeline.

📋 What Zillow property data is publicly available to scrape?

Anyone can open a Zillow listing without logging in and see its price, facts, history and photos — that's exactly what this Actor returns.

Data Category	Publicly Available	Restricted (owner/agent login)
Core facts — address, price, beds/baths, home type, year built	Yes	—
Zestimate, Rent Zestimate, tax history, price history	Yes	—
Listing agent name, brokerage, MLS ID	Yes	—
Schools and nearby comparable homes	Yes	—
Photos, virtual tour and floor plan links	Yes	—
Owner/agent portal status flags and management links	—	Yes
BuyAbility financing pre-qualification data	—	Yes

Zillow Property Scraper only returns publicly visible data — what any visitor sees on the listing page. Nothing behind a login wall.

⚠️ The schema also carries a handful of owner/agent-portal fields — isListingClaimedByCurrentSignedInUser, isCurrentSignedInAgentResponsible, isCurrentSignedInUserVerifiedOwner, propertyUpdatePageLink, moveHomeMapLocationLink, propertyEventLogLink, editPropertyHistorylink, and buyAbilityData. They exist for schema completeness, but since the scraper never authenticates, they read null or false in every row.

🗂️ What data can I extract with Zillow Property Scraper?

Every run returns one record per property covering identity and location, price and quantitative facts, and history/media modules — 186 fields in total, projected onto the same schema whether Zillow exposes the value or not.

Location & identity fields

Field Name	Description
zpid	Zillow's unique property ID
streetAddress / city / state / zipcode / county	Parsed street address components
address	Structured address object
hdpUrl	Canonical Zillow property URL
homeStatus	Listing status (for sale, pending, sold, off-market, etc.) as returned by Zillow
homeType	Property type (single family, condo, townhouse, etc.)
yearBuilt	Year the structure was built
latitude / longitude	Geocoordinates
description / whatILove	Listing description text

Price & quantitative fields

Field Name	Description
price / currency	Current listing price and currency
bedrooms / bathrooms	Bed and bath counts
livingArea / livingAreaValue / livingAreaUnits	Interior square footage
lotSize / lotAreaValue / lotAreaUnits	Lot size
zestimate / rentZestimate	Zillow's automated value and rent estimates
zestimateLowPercent / zestimateHighPercent	Zestimate confidence range bounds
priceHistory	Array of past list/sale price events
taxHistory	Array of past assessed value and tax events
propertyTaxRate / monthlyHoaFee	Tax rate and HOA fee
daysOnZillow / photoCount	Listing age and photo count

History & media modules

Field Name	Description
schools	Array of nearby schools with ratings and distance
nearbyHomes	Array of comparable nearby properties
originalPhotos	Full-resolution photo array
hiResImageLink / mediumImageLink / thumb	Derived photo variants at fixed sizes
virtualTourUrl / interactiveFloorPlanUrl	Virtual tour and floor plan links
listedBy / attributionInfo / brokerageName / mlsid	Listing agent, brokerage and MLS identifiers
mortgageRates / mortgageZHLRates / annualHomeownersInsurance	Financing estimate modules

🔑 How to use Zillow Property Scraper

Zillow Property Scraper runs on the Apify platform — there's no separate signup or API key beyond an Apify account.

1. Open Zillow Property Scraper on its Apify Store listing.
2. Provide the required input — startUrls: one or more Zillow property URLs, bare ZPIDs, or street addresses, one per line.
3. Optionally link searchResultsDatasetId to a dataset from a Zillow search/listing scraper to enrich every property it found, or configure proxyConfiguration to start on a specific proxy tier instead of the automatic no-proxy default.
4. Click Start (or call the Actor via the Apify API).
5. Open the Output tab to browse results across the Overview, Location, Building, Financials, Schools, Nearby Homes, Media and Listing views, or export as JSON, CSV or Excel.

How to scale to bulk property extraction

startUrls is an array (the stringList editor), so a single run accepts as many URLs, ZPIDs or addresses as you paste in — one per line, no separate run per property required. Every property in the list is fetched independently and pushed to the dataset the moment it's ready, so a crash or stop mid-run still leaves you with whatever was already scraped instead of losing the whole batch.

💡 What can you do with Zillow property data?

• 📈 An investment analyst tracking a market uses priceHistory and taxHistory together to spot properties with a widening gap between assessed value and list price.
• 🏗️ A real estate developer scoping a lot uses lotAreaValue, resoFacts and homeType to shortlist parcels before requesting a survey.
• 🏫 A relocation service comparing neighborhoods uses schools and nearbyHomes to build a comparable-homes report for a family evaluating a specific address.
• 🤝 A lead-generation team building outreach lists uses attributionInfo, listedBy and brokerageName to identify which agent or brokerage represents a given property.
• 🤖 An AI agent building a property-valuation tool feeds price, zestimate, livingArea, bedrooms, bathrooms and description into a RAG pipeline so it can answer natural-language questions about a specific address with grounded, structured facts instead of guessing.

🛡️ Why not build this yourself?

Zillow's property page serves its data through a server-rendered __NEXT_DATA__ JSON blob whose wrapper object name changes depending on listing status, plus a separate GraphQL endpoint (ForSaleNonPriorityQuery) and a handful of persisted-query hashes for bonus modules like mortgage rates and home valuations — hashes Zillow rotates periodically, so a hardcoded query breaks silently until you re-discover the new one. On top of that, Zillow's PerimeterX bot defense blocks a plain HTTP client almost immediately; getting past it takes a warmed session and a real browser TLS fingerprint, and when that's still not enough, escalating through paid datacenter and residential proxy tiers. Zillow Property Scraper handles all of that automatically and defaults to no proxy at all, so you only pay for a higher tier when a request is actually blocked.

⛔ How does Zillow Property Scraper handle rate limits and blocking?

Every property starts with no proxy at all. If a request comes back blocked — a non-200 status, a PerimeterX challenge page, or a network error — the shared proxy tier for the whole run escalates: No proxy → Datacenter proxy → Residential proxy, and once Residential is reached, every property in the run stays on it (the tier only ever moves up, never back down, and a block discovered by one property benefits every other property running concurrently). Each property is retried up to six times, with growing delays and a rotated mobile-Safari session fingerprint on every attempt, plus up to three additional retries once already on Residential. If every attempt is exhausted, that property is skipped — no partial or fabricated row is pushed for it — and the run log records the reason (HTTP status, "blocked or empty response", or the exception name) so you can see exactly which inputs failed and why.

⬇️ Input

Zillow Property Scraper takes three input fields, defined in .actor/input_schema.json. Only startUrls is required.

Parameter	Required	Type	Description	Example Value
startUrls	Yes	Array of strings	Zillow property links, bare ZPIDs, or full street addresses — one per line. Bulk input supported.	"https://www.zillow.com/homedetails/7413-First-Landing-Ct-Henrico-VA-23231/72095256_zpid/"
searchResultsDatasetId	No	Array	Dataset ID(s) from a Zillow search/listing scraper, used to enrich every property found there with full details.	"AbCdEfGhIjKlMnOp01"
proxyConfiguration	No	Object	Apify Proxy configuration. Default is no proxy; the run auto-escalates to Datacenter then Residential only if a request is blocked.	{ "useApifyProxy": false }

Example input

{
  "startUrls": [
    "https://www.zillow.com/homedetails/7413-First-Landing-Ct-Henrico-VA-23231/72095256_zpid/",
    "21386 Sunnyside Rd, Los Gatos, CA 95033",
    "24996529"
  ],
  "searchResultsDatasetId": [],
  "proxyConfiguration": {
    "useApifyProxy": false
  }
}

⬆️ Output

Zillow Property Scraper pushes one dataset item per property, every item projected onto the same 186-key schema in the same order — a field Zillow doesn't expose for that particular property comes back as null instead of being left out, so every row has an identical shape. Open the Output tab to browse it split into Overview, Location, Building, Financials, Schools, Nearby Homes, Media and Listing views, or export the whole dataset as JSON, CSV or Excel. The Actor bills on the row_result event, charged once per property that's actually pushed to the dataset — a property that fails every retry produces no row and is never charged.

Example output

{
  "zpid": 72095256,
  "hdpUrl": "https://www.zillow.com/homedetails/7413-First-Landing-Ct-Henrico-VA-23231/72095256_zpid/",
  "homeStatus": "FOR_SALE",
  "streetAddress": "7413 First Landing Ct",
  "city": "Henrico",
  "state": "VA",
  "zipcode": "23231",
  "county": "Henrico County",
  "price": 435000,
  "currency": "USD",
  "bedrooms": 4,
  "bathrooms": 3,
  "livingArea": 2648,
  "livingAreaUnits": "Square Feet",
  "lotSize": 12197,
  "lotAreaUnits": "Square Feet",
  "yearBuilt": 2015,
  "homeType": "SINGLE_FAMILY",
  "zestimate": 441200,
  "rentZestimate": 2450,
  "latitude": 37.4831,
  "longitude": -77.3175,
  "daysOnZillow": 12,
  "photoCount": 38,
  "description": "Beautifully maintained colonial in the sought-after First Landing community...",
  "priceHistory": [
    { "date": "2024-06-01", "event": "Listed for sale", "price": 435000 }
  ],
  "taxHistory": [
    { "year": 2023, "taxPaid": 3210, "value": 398000 }
  ],
  "schools": [
    { "name": "Varina High School", "rating": 4, "level": "High", "distance": 2.1 }
  ],
  "listedBy": [
    { "displayName": "Jane Smith", "phone": "(804) 555-0100" }
  ],
  "brokerageName": "Long & Foster Real Estate",
  "mlsid": "2412345",
  "hiResImageLink": "https://photos.zillowstatic.com/fp/abc123def456-p_f.jpg",
  "virtualTourUrl": null
}

⚙️ How does it work?

Zillow Property Scraper runs each property through a short request pipeline over a session that mimics a real mobile Safari browser at the TLS fingerprint level. It first warms the session against Zillow's homepage, then fetches the property detail page — resolving the ZPID for address or bare-ZPID input in the process — by recursively searching the page's embedded __NEXT_DATA__ JSON for the object carrying that ZPID, rather than a fixed path, so the parser keeps working even when Zillow renames its wrapper objects. A second request hits Zillow's own GraphQL endpoint for the richer modules (price history, tax history, schools, nearby homes, description), and a few best-effort requests fill in bonus modules like mortgage rates when available. Every source is merged, real values always winning over nulls, and projected onto the same fixed schema every run — so your downstream code never has to handle a shifting shape. Only what a logged-out visitor can see is ever returned.

⚠️ The mortgage-rate, home-insights and home-valuation modules are addressed by a persisted GraphQL query hash that Zillow rotates from time to time. When a hash goes stale, those specific fields come back null for that run instead of breaking it — every other field is unaffected.

🔌 Integrations

Zillow Property Scraper works with what you already use — call it directly from your own code, wire it into an AI agent over MCP, or drop it into a no-code automation.

Calling Zillow Property Scraper programmatically

from apify_client import ApifyClient

client = ApifyClient("<APIFY_TOKEN>")
run = client.actor("scraper-engine/zillow-property-details-scraper").call(
    run_input={"startUrls": ["21386 Sunnyside Rd, Los Gatos, CA 95033"]}
)

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["zpid"], item["price"], item["zestimate"])

Works in Go, Ruby, Node.js, cURL — any language that can make an HTTP request.

MCP integration for AI agents

Zillow Property Scraper is reachable through Apify's MCP server at mcp.apify.com. Register it with the Apify CLI — apify mcp install claude-code --tools scraper-engine/zillow-property-details-scraper — or add https://mcp.apify.com?tools=scraper-engine/zillow-property-details-scraper as a custom MCP server URL in Claude Desktop. The CLI installs the same way for Cursor, VS Code, Codex, Kiro and Antigravity, giving any of them the Actor's input schema as a tool and its dataset rows as the tool result.

No-code tools (n8n, Make, LangChain)

In n8n, add the official Apify node (@apify/n8n-nodes-apify) to run this Actor and read its dataset items directly inside a workflow, no manual HTTP Request node needed. In Make, use Apify's app modules — Run an Actor, pointed at zillow-property-details-scraper, then Get Dataset Items — to drop property data straight into a scenario. In LangChain, ApifyWrapper.call_actor() runs this Actor and hands back an ApifyDatasetLoader you can feed directly into a retriever or vector store.

⚖️ Is it legal to scrape Zillow properties?

Scraping publicly available Zillow property listings — addresses, prices, home facts, tax records and photos already published for any visitor to see — is generally lawful and sits under contract and database-rights law rather than data-protection law. Zillow Property Scraper only returns what a logged-out visitor can already see on the listing page, so this is business/product data, and GDPR does not attach to it. The one exception is the listing-agent fields — listedBy, attributionInfo, postingContact and brokerageName — which do carry a named individual's professional contact details; if you store or reuse those at scale, GDPR/CCPA's lawful-basis rules for personal data can apply to that subset. Always review Zillow's Terms of Service for your specific use case, and consult legal counsel if your use case involves bulk storage of personal data.

❓ Frequently asked questions

What Zillow property fields does Zillow Property Scraper return?

The five most-used fields are price, zestimate, bedrooms, bathrooms and homeStatus, alongside 180+ others covering location, history, schools and media — see What data can I extract above for the full breakdown.

Does Zillow Property Scraper require a Zillow account or login?

No. The Actor runs an anonymous session against Zillow's public pages and never authenticates — which is also why the owner/agent-portal fields in the schema always come back null or false.

How many properties can I extract in one run?

As many as you list in startUrls. It's an array field with no fixed cap in the input schema, and each property is processed and pushed to the dataset independently of the others.

What happens if a property is off-market, delisted, or the input doesn't resolve to a real listing?

If Zillow still serves a page for a former listing, homeStatus reflects that directly (Zillow's own status value). If the URL, ZPID or address never resolves to a real property page after every retry, that item is skipped entirely — no row is pushed for it, and the run log records the reason, so a shorter dataset than your input list tells you exactly which inputs failed.

Can I scrape multiple Zillow properties at once?

Yes. startUrls accepts a list — paste as many URLs, ZPIDs or addresses as you want, one per line, and they're all scraped in the same run.

Does Zillow Property Scraper work with Claude, ChatGPT, and other AI agent tools?

Yes. It's reachable through Apify's MCP server (see Integrations above), and any other agent framework can call it as a plain HTTP endpoint through the Apify API.

What's the difference between using a URL/ZPID versus a street address as input?

A URL or bare ZPID goes straight to the property's detail page. A street address is first resolved through Zillow's own search, so the ZPID is only known once the property page loads — it depends on Zillow matching the address to a real listing.

Does Zillow Property Scraper return data in a format LLMs can use directly?

Yes. Every record is typed, normalized JSON with the same field names on every run — no HTML parsing, no CSS selectors. Pass it directly to an LLM, index it into a vector store, or feed it to an agent tool.

What happens when Zillow changes its layout or anti-bot system?

The parser locates the property object by recursively searching for the dict that carries the matching ZPID rather than a fixed JSON path, so it keeps working when Zillow renames a wrapper object, and the output schema stays the same 186 keys regardless. The bonus GraphQL modules addressed by a persisted query hash are treated as best-effort and simply come back empty if Zillow rotates that hash, without affecting the rest of the record.

Can I use Zillow Property Scraper without managing proxies or browser infrastructure?

Yes. The default is no proxy at all, and the Actor only escalates to Apify's Datacenter and then Residential proxy tiers automatically when a request is actually blocked — you never configure sessions, fingerprints, or retries yourself.

Which Zillow property fields work best for AI training data and RAG indexing?

For RAG, index description and whatILove — the highest-information free text in the record. For structured training data, zpid, price, bedrooms, bathrooms, livingArea, homeType, zestimate and yearBuilt return as consistent typed primitives on every record, since every property is projected onto the same fixed schema.

🔗 Related scrapers

Scraper	What it extracts
Website Content Crawler	General-purpose page crawler — pair it with Zillow Property Scraper to pull unstructured pages Zillow doesn't expose as structured fields, like agent bio pages or neighborhood guide content.

💬 Your feedback

Found a bug or missing a field? Let us know — email dev.scraperengine@gmail.com for custom solutions, feature requests, or to report an issue. Active maintenance means the schema and parser get updated as Zillow's site changes.