Redfin Scraper

Scrape Redfin into clean JSON or CSV. It covers homes for sale, recently sold homes, rentals, full property details (Redfin Estimate, tax and price history, schools, climate risk) and housing-market stats. It ships as a CLI and a Python library.

Powered by ScrapeUnblocker. Every page is fetched through the ScrapeUnblocker getPageSource API, which returns Redfin's fully rendered page, so you don't run a browser or manage blocks yourself.

redfin-scraper search "Austin, TX" --beds 3 --max-price 600k --format table
redfin-scraper search 98103 --sold 6mo --format csv -o seattle_sold.csv
redfin-scraper rentals "Denver, CO" --max-price 2500 --type apartment --dogs
redfin-scraper home https://www.redfin.com/WA/Seattle/1624-N-46th-St-98103/home/118803
redfin-scraper market "Boise, ID"

Features

• Homes for sale: price, beds, baths, sq ft, lot size, year built, HOA, property type, status (Active, Coming Soon, Pending, ...), days on market, address, coordinates, neighborhood, open house, badges, key facts, listing remarks, photo count and tour flags.
• Sold homes: --sold 1wk|1mo|3mo|6mo|1yr|2yr|3yr|5yr with the sale date. A sale_price_hidden flag is set when Redfin marks a sale price as not publicly disclosed (in non-disclosure states such as Texas).
• Rentals: houses, condos, townhouses and apartment buildings with rent range, bed/bath/sq ft ranges, available units, specials ("12 WEEKS FREE"), key facts and the leasing phone.
• Full home details (home):
  • Value: list or sold price, Redfin Estimate, rent estimate.
  • Facts: every MLS amenity group (parking, interior, exterior, utilities, ...), key details, county, APN and MLS ID.
  • Records: property history (listed, price changes, sold) and tax history (assessed land and improvement values, taxes per year).
  • Neighborhood: assigned schools with GreatSchools ratings, Walk/Transit/Bike Scores, and First Street climate risk factors (flood, fire, heat, wind, air).
  • Listing: agent and broker, open house, full-size photos.
• Rental details: floor plans with live rent ranges and individual units (unit number, rent, sq ft, available date), unit and community amenities, pet policies with pet rent, parking, office hours and management company.
• Market stats (market): median list and sale price, $/sq ft, sale-to-list %, homes sold and on market, average days on market and offers, year-over-year change, Redfin Compete Score, inventory by bedroom count, and 5 years of monthly history.
• Every Redfin search filter that matters: price, beds, baths, sq ft, lot size, year built, max HOA, stories, property type, listing status, keyword in description, garage, days on market, pets, sort. Any other filter token can be passed straight through.
• Past the 350-home cap: --all splits the price range into bands until every match is collected (or your request budget runs out), deduplicated by property ID.
• Location input your way: "City, ST", a neighborhood name, a ZIP code, or any Redfin search URL pasted from your browser (its filters are kept).
• Robust: retries with backoff on timeouts and transient errors, and detects when Redfin silently ignores your filters instead of returning unfiltered results.
• Output as JSON, JSON Lines, CSV or a terminal table; typed dataclasses in Python.

How it works

Redfin renders pages on the server, and while doing so it calls its own JSON APIs (map search, home-details sections, market trends) and embeds every response in the page (root.__reactServerState.InitialContext). This scraper fetches the page through ScrapeUnblocker's getPageSource endpoint and reads that embedded data directly. It uses no CSS selectors, so it returns the same structured data the page itself is drawn from.

A few Redfin behaviors the scraper handles for you:

• One search returns at most 350 homes. That's Redfin's own cap (page-2 URLs return the same 350). search_all() / --all splits the price range at the median price into bands that each fit under the cap.
• One bad filter drops all filters. If any filter token is unknown, Redfin silently shows the unfiltered search. The scraper compares your filters with the filter string Redfin echoes back and raises FilterRejectedError instead of returning misleading results.
• Place names need a lookup. "Boise, ID" is resolved by typing it into Redfin's own search box with ScrapeUnblocker browser steps (one request). That page is the search result, so an unfiltered search by name costs a single request. ZIP codes and Redfin URLs need no lookup.

Install

git clone https://github.com/ScrapeUnblocker/redfin-scraper.git
cd redfin-scraper
python -m pip install .

Get an API key at scrapeunblocker.com and expose it as SCRAPEUNBLOCKER_KEY (see .env.example):

export SCRAPEUNBLOCKER_KEY=your_key_here

Requires Python 3.9+. The only runtime dependency is the official scrapeunblocker SDK.

CLI usage

redfin-scraper search  LOCATION [filters] [--sold WINDOW] [--sort S] [--all] [-f FORMAT] [-o FILE]
redfin-scraper rentals LOCATION [filters] [--type T] [--dogs] [--cats] [--all] ...
redfin-scraper home    URL [URL ...]
redfin-scraper market  LOCATION [--no-history]
redfin-scraper locate  LOCATION

LOCATION is a place name ("Austin, TX", "Capitol Hill, Seattle"), a ZIP code (78704) or a Redfin search URL or path (https://www.redfin.com/city/30818/TX/Austin/filter/min-beds=3).

Search homes for sale

# 3+ bed houses and townhouses under $600k, newest listings first
redfin-scraper search "Austin, TX" --beds 3 --max-price 600k --type house,townhouse --sort newest -f table

# Every match, not just the first 350 (spends up to 8 requests)
redfin-scraper search "Austin, TX" --beds 3 --all --max-requests 8 -f csv -o austin.csv

# Keyword in the description, garage, max HOA, listed within the last week
redfin-scraper search 78704 --keyword "pool" --garage --max-hoa 300 --max-days 1wk

Sold homes (comps)

redfin-scraper search 98103 --sold 6mo --beds 3 --sort recently-sold -f csv -o comps.csv

Rentals

redfin-scraper rentals 78704 --max-price 2500 --type apartment -f table
redfin-scraper rentals "Denver, CO" --beds 2 --dogs --sort price-asc

Home details, market stats, location lookup

redfin-scraper home https://www.redfin.com/TX/Austin/7500-Lobelia-Dr-78729/home/32797952
redfin-scraper market "Austin, TX" -f json
redfin-scraper market 78704 -f csv          # the monthly history as CSV
redfin-scraper locate "Boise, ID"

Filters

Option	Redfin filter	Applies to
--min-price, --max-price	price (400000, 400k, 1.2m)	sale, sold, rentals
--beds, --max-beds, --baths	bedrooms / bathrooms	sale, sold, rentals
--type	house, condo, townhouse, multifamily, land, manufactured, co-op, other (rentals: house, condo, townhouse, apartment)	all
--status	active, comingsoon, contingent, pending	sale
--sold	1wk, 1mo, 3mo, 6mo, 1yr, 2yr, 3yr, 5yr	sold
--min-sqft, --max-sqft	interior size	sale, sold
--min-lot, --max-lot	lot size in acres	sale, sold
--min-year, --max-year	year built	sale, sold
--max-hoa	HOA dues per month	sale, sold
--min-stories	stories	sale, sold
--keyword	words in the listing description	sale, sold
--garage	has a garage	sale, sold
--max-days	time on market (3d, 1wk, 1mo, ...)	sale
--dogs, --cats	pets allowed	rentals
--sort	recommended, newest, price-asc, price-desc, sqft-desc, recently-sold	all (rentals: first four)
--filter TOKEN	any raw token copied from a Redfin URL (repeatable)	all

Output options: --format json|jsonl|csv|table (default json), --output FILE, --limit N. A summary line, with the result count, Redfin's match count and requests used, goes to stderr so stdout stays clean for piping.

Library usage

from redfin_scraper import Filters, RedfinScraper

scraper = RedfinScraper()  # reads SCRAPEUNBLOCKER_KEY; or RedfinScraper(api_key="...")

# Homes for sale
page = scraper.search("Austin, TX", Filters(min_beds=3, max_price=600_000, sort="newest"))
print(page.region.name, page.total, page.applied_filters)
for home in page.homes:
    print(home.price, home.beds, home.baths, home.sqft, home.address, home.url)

# Past the 350-home cap, within a request budget
everything = scraper.search_all("Austin, TX", Filters(min_beds=3), max_requests=8)

# Sold comps and rentals
sold = scraper.search("98103", sold="6mo", min_beds=3)  # keyword shortcuts work too
rentals = scraper.rentals("78704", Filters(max_price=2500, property_types=["apartment"]))

# Full details of one home (HomeDetails) or rental building (RentalDetails)
details = scraper.home(page.homes[0].url)
print(details.redfin_estimate, details.rent_estimate, details.tax_history[:2])
print([(s["name"], s["rating"]) for s in details.schools], details.climate_risk)

# Market stats with 5 years of monthly history
market = scraper.market("Austin, TX")
print(market.median_sale_price, market.sale_to_list_percent, market.history[-1])

Every result is a dataclass with to_dict(). The parsers (parse_search_page, parse_rentals_page, parse_home_page, parse_market) are public too, so you can feed them HTML you fetched another way.

Errors: RedfinNotFoundError (unknown location or removed home), FilterRejectedError (Redfin ignored the filters; pass check_filters=False to accept the page anyway), RedfinParseError (the page had no data after retries). All derive from RedfinError.

Example output

redfin-scraper search 98103 --beds 3 --max-price 1.2m --type house -f table --limit 5:

price        beds  baths  sqft    property_type  status        address                         city            zip
-----------  ----  -----  ------  -------------  ------------  ------------------------------  --------------  -----
1,199,950    5     2      2610    house          Coming Soon   1624 N 46th St                  Seattle         98103
919,000      4     2      2040    house          Active        9752 Evanston Ave N             Seattle         98103
1,200,000    4     3.5    1785    house          Active        9251 Palatine Ave N             Seattle         98103
799,950      3     2      1280    house          Active        9751 Densmore Ave N             Seattle         98103
599,900      4     1      1500    house          Active        9555 College Way N              Seattle         98103

One home from the JSON output (-f json), trimmed:

{
  "property_id": 118803,
  "url": "https://www.redfin.com/WA/Seattle/1624-N-46th-St-98103/home/118803",
  "status": "Coming Soon",
  "price": 1199950,
  "price_per_sqft": 460,
  "beds": 5.0,
  "baths": 2.0,
  "sqft": 2610,
  "lot_sqft": 4500,
  "year_built": 1922,
  "property_type": "house",
  "address": "1624 N 46th St",
  "city": "Seattle",
  "state": "WA",
  "zip": "98103",
  "latitude": 47.6623241,
  "longitude": -122.3379809,
  "location": "Wallingford",
  "badges": ["EARLY ACCESS", "COMPASS COMING SOON"],
  "key_facts": ["Endless potential", "Corner lot", "Incredible location"],
  "photo_count": 34
}

python examples/home_report.py https://www.redfin.com/ID/Boise/6123-W-Northview-St-83704/home/106669402:

6123 W Northview St, Boise, ID 83704  [Active]
  3.0 bd / 2.0 ba / 1721 sqft, built 1957
  Price: $500,000
  Redfin Estimate: $499,646
  Rent estimate: $2,101/mo
  Property tax 2025: $2,695
  School: Koelsch Elementary School (PreK-6), rating 2/10
  School: Fairmont Junior High School (7-9), rating 3/10
  School: Capital Senior High School (9-12), rating 5/10
  Walk/Transit/Bike: 54/24/73
  Climate risk: flood 1/10, fire 1/10, heat 4/10, wind 1/10, air 8/10
  2026-10-15  Listed                   $500,000
  https://www.redfin.com/ID/Boise/6123-W-Northview-St-83704/home/106669402

redfin-scraper market 78704 --no-history (excerpt):

{
  "region": {"name": "78704", "type": "zipcode", "id": 34098, "url": "https://www.redfin.com/zipcode/78704"},
  "median_list_price": 775000,
  "median_sale_price": 884000,
  "median_sale_price_per_sqft": 516,
  "sale_to_list_percent": 95.7,
  "homes_sold": 46,
  "homes_on_market": 368,
  "avg_days_on_market": 100,
  "compete_score": 32,
  "listings_by_beds": [{"beds": "studio", "active_listings": 3, "median_list_price": 925000}, "..."]
}

(Live data from the day this README was written; your numbers will differ.)

Examples

Script	What it does
examples/search_to_csv.py	Export 3+ bed houses and townhouses in a city to CSV with search_all()
examples/sold_comps.py	Recent sales in a ZIP code with a price-per-sq-ft summary
examples/home_report.py	One-page report: price vs. Redfin Estimate, taxes, schools, Walk Score, climate risk, history

Notes and limits

• Requests: each page fetch is one API request. A search by ZIP or URL is 1 request; a search by place name is 1 (no filters) or 2 (with filters); --all uses up to --max-requests in total.
• 350 per search: without --all, results stop at 350 and the result says so (capped). Narrow the filters or use --all for more.
• Match counts: Redfin reports a total match count (total) for most for-sale and sold searches, but not all; it's null when Redfin didn't send one.
• Disclosure: some fields depend on the listing's MLS. For example, full vs. partial baths and the listing agent aren't always provided, and sale prices are hidden in non-disclosure states.
• Coverage: redfin.com (United States). Requests are routed through the US by default (proxy_country).
• Scrape responsibly: collect public data at a reasonable rate and follow Redfin's terms and the laws that apply to you.

Project layout

redfin-scraper/
├── src/redfin_scraper/
│   ├── __init__.py      # public API + __version__
│   ├── scraper.py       # RedfinScraper: fetch + retry, location lookup, search_all price bands
│   ├── parsing.py       # page data -> Home / Rental / HomeDetails / RentalDetails / Market
│   ├── state.py         # extracts the API responses embedded in Redfin pages; errors
│   ├── urls.py          # Filters -> Redfin filter tokens, search/home URLs
│   ├── models.py        # dataclasses
│   ├── export.py        # JSON / JSONL / CSV / table writers
│   └── cli.py           # `redfin-scraper` command
├── examples/            # runnable scripts
├── tests/               # offline tests (mocked client + trimmed real page data)
├── pyproject.toml
├── requirements.txt
├── Makefile
└── .env.example

Development

make install   # pip install -e ".[dev]"
make lint      # ruff check .
make format    # ruff format .
make test      # pytest -q

The tests run fully offline. They use a fake client and trimmed snapshots of real Redfin page data (agent names and phone numbers replaced), so they cost no API credits. CI runs lint and tests on Python 3.9 and 3.12.

Links

• ScrapeUnblocker: web scraping API that handles blocks, CAPTCHAs and rendering
• Documentation
• Python SDK

License

MIT. Copyright (c) 2026 ScrapeUnblocker.