Zillow Scraper API - Listings, Zestimate, Agents & Cash Flow

Zillow Scraper

Listings, Zestimate, agent contacts, comparables and rental cash flow — in one run

Turn any Zillow search or property page into clean, structured data: prices, photos, agent and broker contacts, valuations, price and tax history, schools, nearby homes and more. Switch on Deals and every property also gets rental cash flow, cap rate, comparable nearby homes and seller-motivation signals — all computed from data already on the page, so they cost no extra requests.

• 🏠 Four workflows — fast search cards, full property records, investment analysis, or known-property lookups.
• 💸 Underwriting built in — cap rate, cash-on-cash, break-even rent and a four-way stress test with assumptions you control.
• 🧾 Export anywhere — JSON, CSV or Excel, through the UI or the API, with flat spreadsheet-friendly aliases next to native Zillow fields.
• 🧭 Honest by design — a field appears only when Zillow actually publishes it; errors go to a separate dataset and are never billed as results.

---

▶️ Your first run

1. Open the Input tab and pick a workflow (Search, Search + details, Deals or Detail).
2. For a search, type a city, ZIP, neighborhood, region slug or full Zillow search URL. For Detail, paste property URLs or ZPIDs.
3. Choose a listing status (For sale, Recently sold, For rent or Any).
4. Set Maximum saved results — start small, inspect the output, then scale up.
5. Click Start, open Output, and export as JSON, CSV or Excel.

{
  "mode": "SEARCH_DETAILS",
  "targets": ["Sonoma, CA"],
  "propertyStatus": "FOR_SALE",
  "maxResults": 20
}

Everything below is optional. The defaults return useful data on the first run.

---

🚀 Pick your goal

Choose one workflow. Search modes take a location; Detail takes properties you already know.

Mode	Best for	What you get
⚡ SEARCH	Scanning a market fast and cheap	Listing cards: price, address, beds/baths, basic features, images, coordinates and source coverage
✨ SEARCH_DETAILS	Complete records for a search	Every search card enriched with the full property page — history, schools, agents, photos, nearby homes
💎 DEALS	Finding and underwriting investments	Full properties plus cash flow, cap rate, comparable nearby homes, valuation gaps, seller signals and a stress test
🏠 DETAIL	Properties you already have	Full records from Zillow URLs, ZPIDs, or a dataset from an earlier run

Search covers for-sale, sold and rental results. Rental search can return building-summary cards; full-detail modes skip those containers and collect individual listings instead of silently expanding an apartment building into every unit.

{
  "mode": "DETAIL",
  "targets": ["15800416"],
  "analyzeDeals": true,
  "maxResults": 1
}

---

✨ What makes the output easy to use

• 🧱 Native + flat aliases. Zillow's original structure is preserved, and spreadsheet-friendly keys such as listingPrice, listingAddress, agent, broker, mls, listingPhotos, listingPriceHistory and listingTaxHistory sit next to it.
• 🔢 Zeros and false are kept. A real 0 favorite count or false garage flag is preserved, not dropped.
• 🚫 No invented data. A field is returned only when Zillow publishes a value; otherwise it is absent and the reason is noted.
• 🗂️ Grouped views. Listings, property details, agents, photos, price history, tax history, schools, deals, stress tests and errors each get their own table view.
• 🧮 Transparent math. Every derived number travels with its assumptions, so a cap rate or stress scenario can be audited, not just trusted.
• 🧯 Separated errors. Source failures and diagnostics land in a dedicated errors dataset and never consume a result charge.

---

📊 Listing & valuation fields

Each property carries identity, pricing, valuation, interest and media fields when Zillow publishes them.

🪪 Identity & location

zpid · propertyUrl · displayAddress · city · state · zipcode · coordinates (latitude/longitude) · homeStatus · homeType

💰 Price & property

unformattedPrice · price · currency · bedrooms · bathrooms · livingArea · lotArea · yearBuilt · parking · published amenities and resoFacts

📐 Valuation

zestimate · rentZestimate · zestimateRangeLow / zestimateRangeHigh · rentEstimateRangeLow / rentEstimateRangeHigh · source range percentages

👀 Interest

pageViewCount · favoriteCount · calculated saveToViewRatioPct

🖼️ Media & history

listingPhotos · mainImage · tours and floor plans · listingPriceHistory · listingTaxHistory · lastSoldPrice · open-house info · nearbySchools

{
  "zpid": "15800416",
  "displayAddress": "17199 Park Ave, Sonoma, CA 95476",
  "unformattedPrice": 1250000,
  "bedrooms": 3,
  "bathrooms": 2,
  "livingArea": 1894,
  "zestimate": 1310000,
  "rentZestimate": 4600,
  "saveToViewRatioPct": 3.1
}

Illustrative shape — real values come from Zillow at run time.

---

👤 Agent & contact fields

When a listing publishes them, you receive normalized contact objects:

• agent — name, phoneNumber, email, licenseNumber, coAgentName
• broker — name, phoneNumber
• mls — id, name, lastChecked, lastUpdated

Only contacts actually published for a listing are returned. Zillow's promoted contact forms are not treated as verified listing-agent details, so a phone or email can legitimately be missing on some records. Use the flat agent / broker / mls aliases for a clean CSV export.

---

💸 Cash flow & stress test

Enable Deals (or analyzeDeals) and every purchasable listing with an asking price and a Rent Zestimate receives four scenarios — no separate analysis service required.

Scenario	Rent	Mortgage rate
Base	Zillow Rent Zestimate	Your input, the Zillow rate, or the documented fallback
Lower rent	10% below estimate	Base rate
Higher rate	Base estimate	Base rate + 2 pts
Combined stress	10% below estimate	Base rate + 2 pts

Key fields: analysis.capRatePct, analysis.monthlyCashFlowUsd, analysis.cashOnCashPct, analysis.grossYieldPct, analysis.breakEvenRentUsd, analysis.scenarios, stressMonthlyCashFlowUsd, stressCashFlowPositive, debtServiceCoverageRatio, rentBufferPct.

{
  "mode": "DEALS",
  "targets": ["Sonoma, CA"],
  "maxResults": 10,
  "dealSettings": {
    "downPaymentPct": 25,
    "interestRatePct": "7",
    "vacancyPct": 5,
    "maintenancePct": 5
  }
}

The model also uses the loan term, insurance rate and closing-cost assumptions. Tax uses a published annual amount when available, otherwise price × the source tax rate; it does not predict reassessment after purchase. Missing taxes or HOA are flagged in notes and modelInputsComplete.

Rental listing prices are monthly rent, so they are never used as acquisition prices. sellerEquityUsd stays null because outstanding debt is unknown. Zestimate and Rent Zestimate are estimates, not guaranteed sale prices or achievable rents, and seller and demand scores are rules-based research signals, not predictions of intent.

---

🏘️ Comparable nearby homes

When Zillow's property page carries nearby homes, Deals adds an analysis.comparables block built from that same payload — no extra request.

Field	Meaning
count, statusCounts	How many comparables, split by sold / for-sale / other
medianPriceUsd, medianPricePerSqft	The block's median price and $/sqft
subjectPricePerSqft	This property's own $/sqft
subjectVsComparablesPct	How far above (+) or below (−) the block this property is priced per sqft
comparables[]	Each nearby home: address, status, price, beds, baths, living area, $/sqft

The flat alias comparablePriceDeltaPct mirrors subjectVsComparablesPct for spreadsheets. Comparables are Zillow nearby homes for quick context, not a formal appraisal.

---

📈 Market report

With analysis enabled, the run writes a self-contained HTML market report to the key-value store (market-report). Beyond the medians it shows:

• 💵 Price and $/sqft spread — P25–P75 quartiles, not just a midpoint.
• 📏 Median price-vs-Zestimate across the run.
• 🏷️ The property-type mix of what you collected.
• ⏳ A days-on-market breakdown (fresh / recent / aging / stale).
• 🔥 A ranked table of the best yields or the most motivated sellers in the run.

It summarizes up to 2,000 saved properties and is an orientation tool, not an exhaustive market valuation.

---

📥 Simple input

The form has four visible controls:

Field	What to enter
mode	SEARCH, SEARCH_DETAILS, DEALS or DETAIL
targets	One target per line: locations/search URLs for search modes, property URLs/ZPIDs for Detail
propertyStatus	For sale, for rent or recently sold for location searches; leave Any to use the status already in a pasted URL
maxResults	Maximum successful unique results for the whole run, across all targets; default 50

The simpler form keeps the existing output. Full property, contact, history, media, school, comparable and analysis fields remain available in their corresponding workflows. Search stays the quick listing-card workflow; choose Search + details for the complete property record, or Deals for the same data plus the analysis layer.

Locations and links

Search targets can be cities (Sonoma, CA), ZIPs (90210), neighborhoods, region slugs, or Zillow search URLs. A pasted search URL preserves the filters and map area already set on Zillow. The listing-status selector applies to plain locations; direct property lookups retain the actual source status.

Internal price, beds/baths, HOA, age, photo, yield and seller-signal result filters, plus custom result sorting, have been removed. Old requests with active filters receive a clear input diagnostic before collection starts. Neutral legacy defaults are accepted. To narrow a search, configure it on Zillow and paste its search URL.

Result limit, pagination and immediate output

maxResults applies globally after duplicate removal. A failed, duplicate or unsupported container card does not consume a saved-result slot. Results stream as each requested record finishes; the Actor processes a page before requesting more instead of waiting for all searches to finish. Collection stops at the saved-result or run-charge limit, source exhaustion, or an explained source limit.

The Actor follows up to 20 visible result pages per location automatically when needed. Zillow can expose fewer listings than its reported total, and a broad location can hit its visible-page cap. OUTPUT distinguishes a result limit, page limit and source cap. For broader coverage, add smaller ZIP or neighborhood searches. A limit of 10,000 does not imply that one location exposes 10,000 listings.

Existing API integrations and model settings

Legacy searchQueries, startUrls, zpids and datasetId remain supported through the API. Use either targets or the legacy target fields, rather than supplying both. datasetId remains an optional hidden Detail input with read permissions, so existing dataset-to-detail workflows still work. Explicit legacy maxPagesPerSearch caps are honored.

Deal calculations keep their existing assumptions. API callers may set analyzeDeals: true for Search + details or Detail, and may pass dealSettings containing assumptions such as downPaymentPct, interestRatePct, loanTermYears, vacancyPct, maintenancePct, insuranceRatePct and closingCostPct. These settings change calculations, not source collection or result filtering. Every calculated scenario retains its assumptions in the output.

---

📦 Outputs & table views

The run writes several datasets and one report:

Destination	Contents
Results (default dataset)	Compact overview and valuation view of every saved property
Property details & contacts	Facts, agents, photos, price/tax history and schools
Deal analysis	Cap rate, cash flow, cash-on-cash, comparables delta and seller signals
Stress test	Base vs stressed cash flow, DSCR and rent cushion
Errors	Unavailable sources and diagnostics, separate from paid results
Market report (key-value store)	The HTML summary described above
OUTPUT / Run summary	Saved counts, effective options, source coverage and stop reasons

Grouped views are included copies of the same records; they do not add result charges.

---

📤 Deal analysis example

{
  "displayAddress": "17199 Park Ave, Sonoma, CA 95476",
  "unformattedPrice": 1250000,
  "analysis": {
    "capRatePct": 4.1,
    "monthlyCashFlowUsd": -320,
    "cashOnCashPct": -1.2,
    "pricePerSqft": 660,
    "priceVsZestimatePct": -4.6,
    "motivatedSellerScore": 38,
    "comparables": {
      "count": 6,
      "medianPricePerSqft": 705,
      "subjectVsComparablesPct": -6.4
    },
    "scenarios": [
      { "scenario": "base", "monthlyCashFlowUsd": -320 },
      { "scenario": "combined_stress", "monthlyCashFlowUsd": -980 }
    ]
  },
  "comparablePriceDeltaPct": -6.4
}

Illustrative — values depend on the live listing and your assumptions.

---

🔌 API examples

Run the Actor with your token and retrieve the dataset. Replace YOUR_TOKEN.

cURL — start a run

curl -X POST "https://api.apify.com/v2/acts/trakk~zillow-scraper/runs?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"mode":"SEARCH","targets":["Austin, TX"],"maxResults":25}'

cURL — run synchronously and get the dataset

curl -X POST "https://api.apify.com/v2/acts/trakk~zillow-scraper/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"mode":"DEALS","targets":["Austin, TX"],"maxResults":10}'

Python

from apify_client import ApifyClient

client = ApifyClient("YOUR_TOKEN")
run = client.actor("trakk/zillow-scraper").call(run_input={
    "mode": "SEARCH_DETAILS",
    "targets": ["Austin, TX"],
    "maxResults": 20,
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["displayAddress"], item.get("unformattedPrice"))

The same JSON input works from the Apify Console, the API, the JavaScript/Python clients and any MCP-compatible client.

---

⚙️ Following a run

• 📊 Run summary (OUTPUT) reports saved counts, effective limits, source coverage and why collection stopped.
• 🧾 Errors dataset separates source failures from real results.
• 🧮 Charge protection: set a maximum run charge in the Console; the Actor checks it before more work and stops when another result cannot be charged.
• 🗺️ Coverage fields distinguish a full page sweep, your page limit, a source error and Zillow's own paging limit — an empty run is explained, never guessed.

---

💲 Pricing

The result-event rate is $0.002 per saved listing ($2 per 1,000), including Search, Search + details, Deals and Detail. Stress scenarios, comparables and source-backed compatibility fields are included at no extra charge.

Only useful successful records saved to Results receive a result charge. Source failures, invalid inputs and diagnostic records are not sold as property results. Removed filters do not trigger an excluded-property analysis charge. Set the maximum run charge in Apify to bound result spending; the live Pricing tab is authoritative.

---

❓ FAQ

Which mode should I start with?

Use Search to scan a market cheaply, Search + details for complete records, Deals for investment analysis, and Detail when you already have URLs or ZPIDs.

Do I need the Deals mode for cash flow?

Yes for the automatic always-on analysis. You can also set analyzeDeals: true in Search + details or Detail. The fast Search mode has no detail page to analyze.

Are agent phone numbers and emails always included?

No. Only contacts a listing actually publishes are returned. Zillow's promoted contact forms are not treated as verified agent details.

What are comparables based on?

The nearby homes Zillow already returns on the property page. They give quick $/sqft context versus the block; they are not a formal appraisal.

Can a successful run return no results?

Yes. Invalid input, unavailable properties and empty source searches each have a separate explanation in the run summary. No placeholder listing is ever created.

How is rent handled for rental listings?

A rental price is monthly rent, so it is never used as a purchase price. Income metrics are unavailable for those records.

Why is a field missing on some properties?

Because Zillow did not publish that value for that listing. The Actor never fills a gap with an invented number.

Can I reuse a previous run as input?

Yes. Through the API, use mode: "DETAIL" and datasetId for the earlier dataset; records must contain a zpid, detailUrl, hdpUrl or url. The compact form accepts property URLs and ZPIDs in Targets.

Will removed filters still charge for excluded properties?

No. Active legacy result filters are rejected before source collection. This release does not charge excluded-property-analysis. A search narrowed on Zillow returns its matching source listings directly.

Which locations are supported?

Any city, state, county, ZIP or neighborhood Zillow can resolve, plus full Zillow search URLs and region slugs. Results reflect Zillow’s available US property coverage.

What export formats are available?

JSON, CSV and Excel from the Output tab, or any format via the API.

Is this affiliated with Zillow?

No. This is an independent scraper for publicly available data. Zestimate and Rent Zestimate are Zillow estimates, not guaranteed values.