Scrape · Zillow
Every pin is a record.
Zillow is a map of listings. Fetch turns each one into structured data: price, beds, baths, sqft, Zestimate and more.
Search the map, not the site.
Point Fetch at a city, a ZIP, or a drawn map region. For-sale, sold, or rentals, the response comes back as the same typed listings.
- By city
zillow.com/{city}-{state}/
- By ZIP
zillow.com/homes/{zip}_rb/
- Map bounds
?searchQueryState={bounds}
- Status
for-sale · sold · rentals
Watch the price move.
Every listing carries its full price history. Scrape the events, not just today's number, and see how a home really priced.
Zestimate, list, and sold.
The three numbers that never agree, side by side for every property, so your model can learn the gap.
Zestimate
$761K
Zillow's model estimate.
List price
$749K
What the seller asked.
Last sold
$742K
What it closed at.
Zillow does not want a bot.
Between the map and the data sits a stack of defenses. Fetch clears each one so you never see them.
Bot walls
PerimeterX and CAPTCHA gates trip on the first burst. A real browser from rotating egress clears them.
Lazy map tiles
Listings load as you pan and scroll the map. The page renders fully, then every card is read.
Detail behind clicks
Price history and facts open in panels and tabs. They are triggered and captured for you.
Rate limits
Repeat hits from one IP get throttled fast. Each call runs from fresh egress, so volume holds.
Own the whole market.
One address or an entire metro, the shape stays the same. Structured Zillow data, without the parsing or the blocks.
What teams build with it