Zillow Scraper Pro — Zillow API with Built-In MCP Server for AI Agents

The Zillow scraper built for AI agents, real estate investors, and data teams. Extract Zillow property listings, Zestimates, price history, tax records, school ratings, and agent contact info from any ZIP code, city, or Zillow search URL — with a built-in MCP server that plugs straight into Claude, GPT, Cursor, and other LLM agents.

Unlike every other Zillow scraper on Apify, this one is designed to be called by AI — not just by humans. It ships with a full JSON Schema, a two-tier pricing model (fast summary vs. deep detail), and anti-bot resilience tuned for Zillow's Imperva / PerimeterX WAF.

Use cases: real estate lead generation, MLS-style listing monitoring, Zestimate data extraction, property investor comp analysis, rental market research, AI-powered real estate agents, bulk Zillow data export, house price history tracking.

---

Contents

• Why Zillow Scraper Pro
• What it scrapes from Zillow
• Built for AI agents (MCP-native)
• Anti-bot: why this one doesn't break
• Input
• Output schema
• Pricing — two-tier pay-per-event
• Usage examples
• Deployment
• Troubleshooting
• FAQ
• Changelog

---

Why Zillow Scraper Pro

Feature	Zillow Scraper Pro	Typical Apify Zillow Scrapers
MCP server built in	✅ Native Claude / LLM integration	❌
Two-tier pricing	✅ Fast summary + deep detail	❌ One price fits all
Zestimate transparency	✅ zestimateStatus + comp-based fallback for new construction	❌ Null zestimates with no context
Anti-bot resilience	✅ TLS fingerprinting + pre-warmed session pool	⚠️ Breaks on every Zillow update
Rent Zestimate	✅ Included	⚠️ Sometimes
School ratings + walk score	✅ Included in detail mode	⚠️ Usually extra cost
ZIP code, URL, and city search	✅ All three	⚠️ Usually one
For Sale / For Rent / Recently Sold	✅ All three	⚠️ Usually for-sale only

---

What it scrapes from Zillow

Search inputs:

• Scrape Zillow by ZIP code — e.g. ["10001", "90210", "78701"]
• Scrape Zillow by city or neighborhood URL — e.g. https://www.zillow.com/austin-tx/
• Scrape Zillow by custom filtered search URL — price, beds, sqft, property type, etc.
• Listing types: for_sale, for_rent, recently_sold

Data extracted per property (detail mode):

• Pricing: list price, Zestimate, Rent Zestimate, price/sqft, price history, tax assessed value
• Property facts: beds, baths, sqft, lot size, year built, property type, stories, HOA fee
• Location: full address, ZIP code, latitude, longitude
• Real estate context: days on market, listing status, MLS name, MLS ID
• Agent & broker: agent name, phone, broker name, broker phone
• School data: school name, rating (GreatSchools), grades served, distance, public/private
• Walkability: Walk Score, Transit Score, Bike Score
• Tax history: year-by-year tax paid and assessed value
• Rich media: hi-res photo URLs, floor plans where available
• Property description: full listing description text
• resoFacts: appliances, flooring, cooling, heating, parking, and every structured field Zillow publishes

---

Built for AI agents (MCP-native)

Most Zillow scrapers output JSON and stop there. Zillow Scraper Pro publishes a full Model Context Protocol schema so LLM agents (Claude, GPT, Cursor, LangChain, LlamaIndex) can reason about the data before calling it.

Three ways to use with AI

1. Via the Apify MCP server (no code)

Server URL: https://mcp.apify.com/mcp

Add to Claude Desktop, then ask naturally: "Find 3-bed houses under $500K in Austin listed in the last 30 days." Claude calls the actor, filters the results, and answers. No code, no SDK.

2. Via the built-in local MCP server

$node src/mcp-server.js

Runs a standards-compliant MCP server on your machine, exposing search_zillow, get_property_details, and bulk_scrape_zip as callable tools. Connect any MCP client.

3. Via the Apify API Same actor, standard REST/JS client. Use when you want a traditional scraper API without the AI layer.

Agent-friendly output fields

Every record includes three fields specifically for agent reasoning:

Field	Purpose
dataDepth	summary | detail | summary_fallback — tells the agent whether to re-request with dataDepth: "detail"
zestimateStatus	available | new_construction | not_published — explains why a Zestimate is null instead of hiding it
estimatedValuePPSF	Comp-based valuation fallback computed from the median price-per-sqft in the same search. Useful when Zillow won't publish a Zestimate (new construction, coops, luxury).
scrapedAt	ISO 8601 — lets the agent assess data freshness
source	Always "zillow" — for multi-source pipelines

---

Anti-bot: why this one doesn't break

Zillow runs Imperva / PerimeterX with TLS fingerprinting, JavaScript canvas/WebGL fingerprinting, navigator.webdriver detection, and behavioral heuristics. Most Zillow scrapers on the market break within hours of a Zillow deploy. This actor uses:

1. HTTP-only stealth with Chrome TLS fingerprinting (got-scraping) — sub-second requests, no browser overhead, no JavaScript runtime to fingerprint
2. Residential proxy routing via Apify's residential pool — datacenter IPs are blocked at the edge
3. Pre-warmed session pool with session affinity — survives rate-limit bursts that kill fresh-session scrapers
4. __NEXT_DATA__ + gdpClientCache extraction — pulls data from Zillow's embedded JSON payloads, so layout changes don't break selectors
5. Auto-retry on PerimeterX challenges — detects challenge pages and retries with a fresh residential IP (up to 4 retries, with exponential backoff)
6. Configurable human-like delays — 3–8 second jitter between requests (tunable per run)

The result: sub-minute runs at 20-property scale, and weeks between block incidents instead of hours.

---

Input

Core fields

Field	Type	Default	Description
location	string[]	—	Required. ZIP codes ("10001"), city names ("Austin, TX"), or any Zillow URL. Mix freely.
listingType	string	for_sale	for_sale, for_rent, or recently_sold
dataDepth	string	detail	summary — fast, cheap (search-page data only). detail — rich, visits each listing page. auto — tries detail, falls back to summary on repeated failures.
maxResults	integer	100	Max properties to collect per run
filters	object	{}	Pre-scrape filters (see below) — reduces cost by skipping non-matching listings before detail fetches

Filters object

Filter	Type	Description
minPrice	integer	Minimum list price
maxPrice	integer	Maximum list price
minBeds	integer	Minimum bedrooms
maxBeds	integer	Maximum bedrooms
minBaths	number	Minimum bathrooms
maxDaysOnMarket	integer	Only listings on market ≤ N days
propertyTypes	string[]	e.g. ["single_family", "condo", "townhouse"]

Advanced (optional)

Tucked under an advanced object so they don't clutter the main input. You rarely need to touch these.

Field	Type	Default	Description
advanced.proxyConfig	object	RESIDENTIAL	Apify proxy settings — datacenter proxies will be blocked by Zillow
advanced.maxConcurrency	integer	3	Parallel requests — keep 1–5 for Zillow safety
advanced.requestDelayMs	integer	4000	Delay between requests in ms (3000–8000 recommended)

Minimum input

One entry in location — that's it.

Examples

Investor comp pull — 3BR homes under $600K in two Austin ZIP codes:

{
  "location": ["78701", "78702"],
  "listingType": "for_sale",
  "dataDepth": "detail",
  "maxResults": 50,
  "filters": {
    "minBeds": 3,
    "maxPrice": 600000,
    "propertyTypes": ["single_family", "townhouse"]
  }
}

Rental market survey — city name, summary only (fast and cheap):

{
  "location": ["Austin, TX"],
  "listingType": "for_rent",
  "dataDepth": "summary",
  "maxResults": 200
}

Direct Zillow search URL (paste any filtered search from your browser):

{
  "location": ["https://www.zillow.com/austin-tx/homes/3-_beds/?searchQueryState=..."],
  "listingType": "for_sale",
  "dataDepth": "detail",
  "maxResults": 100
}

AI agent call via MCP (no JSON needed — Claude writes this for you):

"Find 3-bed houses under $500K in Austin listed in the last 30 days."

---

Output schema

Each dataset item is one property. dataDepth tells you which tier the item came from.

Summary fields (all runs, dataDepth: "summary")

{
  "zpid": "12345678",
  "url": "https://www.zillow.com/homedetails/...",
  "address": "123 Main St, New York, NY 10001",
  "price": 850000,
  "beds": 3,
  "baths": 2,
  "sqft": 1400,
  "propertyType": "condo",
  "listingType": "FOR_SALE",
  "listingStatus": "For Sale",
  "daysOnMarket": 12,
  "zestimate": 860000,
  "zestimateStatus": "available",
  "rentZestimate": 4200,
  "estimatedValuePPSF": null,
  "latitude": 40.7128,
  "longitude": -74.006,
  "yearBuilt": 1985,
  "imgSrc": "https://...",
  "scrapedAt": "2026-04-15T12:00:00.000Z",
  "source": "zillow",
  "dataDepth": "summary"
}

Detail fields (dataDepth: "detail")

All summary fields plus:

{
  "lotSizeSqft": 2500,
  "stories": 3,
  "description": "Bright corner unit...",
  "monthlyHoaFee": 450,
  "taxAssessedValue": 720000,
  "pricePerSqft": 607,
  "priceHistory": [
    { "date": "2023-03-15", "price": 795000, "event": "Listed for sale", "pricePerSqft": 568 }
  ],
  "taxHistory": [
    { "year": 2023, "taxPaid": 9800, "value": 700000 }
  ],
  "schools": [
    { "name": "PS 1 Tottenville", "rating": 7, "grades": "K-5", "distance": 0.3, "type": "public" }
  ],
  "agent": {
    "agentName": "Jane Smith",
    "agentPhoneNumber": "212-555-0100",
    "brokerName": "Compass",
    "mlsName": "REBNY",
    "mlsId": "REBNY-123456"
  },
  "walkScore": 98,
  "transitScore": 100,
  "bikeScore": 85,
  "resoFacts": {
    "appliances": ["Dishwasher", "Refrigerator"],
    "flooring": ["Hardwood"],
    "heating": ["Forced air"],
    "cooling": ["Central"]
  },
  "photos": ["https://..."],
  "dataDepth": "detail"
}

---

Pricing — two-tier pay-per-event

This actor uses pay-per-event (PPE) pricing. You pay only for what you actually use, and you get two distinct tiers so you never overpay for depth you didn't need.

Event	Price	When you pay	Use when
apify-actor-start	$0.00005 (first 5 seconds waived)	Once per run — Apify platform event, effectively free for almost all runs	Automatic
search-page-processed	$0.002	Per ZIP / search URL handled	Always — covers the anti-bot cost of the search endpoint
property-summary-returned	$0.003	Per property with ~20 core fields (address, price, Zestimate, beds, baths, sqft, yearBuilt, agent, photos)	Monitoring, alerting, bulk exports — when you need breadth over depth
property-detail-enriched	$0.015	Per property with 40+ fields (full description, resoFacts, HOA, tax, walk/transit scores, price history, schools)	Comp analysis, lead generation, due diligence — when you need depth over breadth

Example costs:

• Daily monitoring sweep of 200 ZIP codes × 50 properties summary = $30.40 / day → $912 / month
• Weekly deep comp analysis of 100 properties with full detail = $1.50 / week → $6 / month
• One-time market report: 500 properties detail-mode = $7.50

Compare to single-tier competitors: typical detail-scrapers charge $0.003–$0.005 per property regardless of depth, so you pay the same for shallow fields as for the full resoFacts + schools + tax stack. With Zillow Scraper Pro you only pay the premium when you actually need the premium data.

---

Usage examples

Scrape all Zillow for-sale listings in a ZIP code

{
  "location": ["10001"],
  "listingType": "for_sale",
  "dataDepth": "detail",
  "maxResults": 200
}

Monitor a city in summary mode (cheapest possible)

{
  "location": ["https://www.zillow.com/new-york-ny/homes-for-sale/"],
  "listingType": "for_sale",
  "dataDepth": "summary",
  "maxResults": 500
}

Nightly rental-market sweep across multiple ZIPs

{
  "location": ["90210", "90036", "90024"],
  "listingType": "for_rent",
  "dataDepth": "detail",
  "maxResults": 300,
  "advanced": {
    "requestDelayMs": 5000
  }
}

Recently sold comps for an investor's target ZIP

{
  "location": ["78701"],
  "listingType": "recently_sold",
  "dataDepth": "detail",
  "maxResults": 100
}

---

Deployment

Prerequisites

• Apify CLI: npm install -g apify-cli
• Apify account with residential proxy access

Steps

cd zillow-scraper-pro
npm install
apify login
apify run        # local test
apify push       # deploy to Apify Store

Local testing input

Create .actor/input.json:

{
  "location": ["10001"],
  "listingType": "for_sale",
  "dataDepth": "detail",
  "maxResults": 10,
  "advanced": {
    "requestDelayMs": 3000
  }
}

---

Troubleshooting

Empty results? Zillow may have shipped a change. Patches ship within 48 hours of a confirmed break — file an Issue on the actor page.

Getting blocked? Confirm advanced.proxyConfig is set to RESIDENTIAL. Datacenter IPs are blocked at Zillow's edge immediately.

Null zestimate? Check zestimateStatus. It will be one of:

• available — Zillow published a value (in the zestimate field)
• new_construction — property built in the last 5 years; Zillow doesn't publish Zestimates for these. Use estimatedValuePPSF as a comp-based fallback.
• not_published — coops, mixed-use, or luxury listings where Zillow itself declines to publish. Fallback also available when sqft is known.

Slow runs? advanced.requestDelayMs of 3000–8000ms is the safe range for Zillow. Lowering below 2000ms significantly increases block rates.

Running into the $5 Apify free-tier limit? Start with dataDepth: "summary" and a small maxResults to explore the schema. Detail mode is designed for production use where the depth justifies the per-property price.

---

Frequently asked questions

Q: How is this different from maxcopell/zillow-scraper? A: That scraper is search-only at $0.002/property; no detail pages, no MCP, no Zestimate transparency. Zillow Scraper Pro gives you both tiers in one actor plus AI-agent integration.

Q: Does this work for Zillow rentals? A: Yes. Set listingType: "for_rent". Both Zestimate and Rent Zestimate are extracted in detail mode.

Q: Can it handle recently sold properties? A: Yes. Set listingType: "recently_sold" to pull sold comps for investor analysis.

Q: Does this bypass Zillow's login wall? A: This actor only scrapes publicly accessible Zillow pages. It does not access logged-in-only data, scrape personal information beyond what Zillow publishes, or violate authenticated-only features.

Q: Is scraping Zillow legal? A: Web scraping public pages is generally permitted under the hiQ v. LinkedIn precedent for publicly available data, but you are responsible for complying with Zillow's Terms of Service and applicable laws in your jurisdiction. This actor is provided for legitimate research, due diligence, and AI tool-use scenarios.

Q: Can I use this commercially? A: Yes. Output is yours; usage is subject to Apify's terms and Zillow's ToS. The actor source is Apache-2.0 licensed.

---

Changelog

v3.1 (April 2026) — Major performance + link-correctness pass. Sub-minute ZIP scrapes at 50-property scale, correct city-name URL handling for for_rent and recently_sold, pagination query-string safety (prevents duplicate searchQueryState params), direct Zillow URL inputs now pass through unchanged instead of being re-slugged, proxy pool size reduced for faster warm-ups, and detail fetches no longer race three proxies in parallel (cuts bandwidth ~2/3 with no measurable hit rate loss).

v2.9 (April 2026) — Added zestimateStatus transparency field, estimatedValuePPSF comp-based fallback, rentZestimate extraction, tiered pay-per-event pricing, built-in MCP server.

---

License

Apache 2.0

---

Related searches

Common questions and workflows this actor is built to serve:

• How do I scrape Zillow listings in bulk? → Set location to a list of ZIP codes and dataDepth: "summary" for the cheapest sweep.
• Is there a Zillow API I can use with Claude or GPT? → Yes — this actor ships with a built-in MCP server, usable from Claude Desktop, Cursor, or any MCP client.
• How do I extract Zestimates and Rent Zestimates programmatically? → Both are first-class fields in detail mode. zestimateStatus tells you why a value is missing (new construction, coop, not published) so you can fall back to estimatedValuePPSF.
• How do I get Zillow recently sold comps? → Set listingType: "recently_sold" with the target ZIP or city.
• How do I pull Zillow rental listings for a market? → Set listingType: "for_rent" with ZIPs or a city name like "Austin, TX".
• Does this extract school ratings and walk scores? → Yes, both appear in dataDepth: "detail" runs.
• How do I get Zillow price history and tax history? → Both are structured arrays (priceHistory, taxHistory) in detail mode.
• Can I feed a filtered Zillow search URL directly? → Yes — paste any Zillow URL (ZIP path, city slug, or full searchQueryState URL) into location and the actor passes it through unchanged.
• How do I monitor a ZIP code for new listings nightly? → Schedule a summary-mode run with filters.maxDaysOnMarket: 1.
• How does this compare to other Apify Zillow scrapers? → See the Why Zillow Scraper Pro comparison table.