🚗 Kelley Blue Book Scraper

TL;DR

Pull live used-car listings from Kelley Blue Book by make, model, year range, ZIP, and search radius. Every record carries the KBB Price Advisor grade (GREAT / GOOD / FAIR / HIGH), KBB's own deal-quality flag, so you can filter for underpriced inventory in one query. Built for used-car flippers hunting margin in their metro, dealer competitive-intel desks, fleet sourcing, insurance and warranty underwriters, and AI shopping agents that need clean validated VINs and dealer geo. Sub-minute runs, ~$0.36 per 100 listings on the GOLD plan.

---

Try it on a small dataset, then let us know what you think in a review.

---

What does Kelley Blue Book Scraper do?

Today, a used-car flipper hunting margin in Atlanta opens kbb.com 50 times per day, types in different make / model / year / ZIP combinations, eyeballs every listing for "is this priced below KBB's Fair Purchase band?", and copies the few candidates into a spreadsheet to call dealers about. A dealer competitive-intel analyst at a 12-store group does the same thing weekly across every neighbor dealership. An insurance underwriter pulls valuations one VIN at a time off the public KBB site to feed actuarial models. An AI shopping agent that wants to answer "find me a 2022 Tacoma in TX under $30K" has nowhere to get structured KBB data.

This actor turns all of that into one API call. A daily monitor on a make + model + ZIP combo finishes in under a minute, costs well under $1, and lands in your warehouse with dealer geo, days-on-site, and the KBB deal grade already joined per listing.

The killer field is kbbPriceAdvisor.kbbGrade. KBB itself rates every listing it tracks as GREAT, GOOD, FAIR, or HIGH deal, that's the same grade you see as a green / yellow / red badge on the consumer site. Filter your dataset on kbbGrade == 'GREAT' and you have your shortlist of underpriced inventory in the metro you searched. None of the existing free KBB scrapers surface this field.

For AI agents and LLM pipelines, every record also carries a 300-500 character agentMarkdown summary card you can paste straight into a Claude or ChatGPT context, plus a fieldCompletenessScore so agents can self-filter sparse rows.

Why scrape Kelley Blue Book?

KBB is the dominant US used-car valuation reference. Dealers price against KBB's Fair Purchase band, consumers shop against the same numbers, lenders and insurers reference the same bands for their models. KBB's killer field is the deal grade, GREAT / GOOD / FAIR / HIGH, that lets a buyer filter underpriced inventory in a single query. Nobody else publishes that signal at this depth.

Manual workflow killed: opening kbb.com 50 times a day, retyping make + model + ZIP, eyeballing every listing for "is this below the Fair Purchase band?". A used-car flipper finds roughly one $2K-margin underpriced car per 200 listings scanned, so the per-record cost pays back on the first flip. Insurance underwriters, dealer competitive-intel desks, and AI shopping agents all need the same VIN-validated, geocoded feed, and the free KBB scrapers on the Store do not surface the deal grade at all.

Who needs this?

• Used-car flippers and wholesalers sweeping a 100-mile radius daily for kbbGrade: 'GREAT' listings to flip in the same week
• Dealer competitive-intelligence desks monitoring their five nearest competitors' inventory and pricing daily across make / model
• Insurance and warranty underwriters pulling VIN-validated valuations to train and update actuarial models
• Fleet ops and corporate procurement sourcing the best-priced examples of a fleet vehicle across a wide radius
• Auto-finance and lending platforms building real-time valuation models with VIN-decoded vehicle squibs and live KBB Fair Purchase bands
• Used-car market analysts and journalists sizing local supply, pricing distributions, and inventory turnover by metro
• AI shopping agents and natural-language car search that need a clean, validated, agent-ready stream of KBB listings as a tool result

How to use Kelley Blue Book Scraper

1. Open the Input tab on the actor page
2. Type your make (e.g., BMW) and optionally a model (e.g., 3 Series)
3. Set zip to the 5-digit US ZIP at the center of your search and choose a searchRadius in miles
4. Optionally narrow with yearMin, yearMax, priceMax, and mileageMax
5. Pick condition: USED (default), NEW, or CERTIFIED
6. Set maxItems (default 10, conservative for the daily auto-test; raise for production runs)
7. Click Start or call via the Apify REST API or CLI

You can run on demand from the Console, schedule daily monitoring with Apify's cron scheduler, integrate with Make, Zapier, n8n, Slack, Google Sheets, or pipe directly into a Claude or ChatGPT custom agent.

How much will scraping Kelley Blue Book cost?

This actor is priced per event:

• Actor Start: $0.005 once per run
• KBB listing record: tiered, charged per record written

Apify plan	$/1000 records
FREE	$5.00
BRONZE	$4.50
SILVER	$4.00
GOLD	$3.50
PLATINUM	$3.00
DIAMOND	$3.00

A typical 100-listing search on the GOLD plan is roughly $0.36 plus a few cents of Apify residential proxy bandwidth. Daily monitoring of one make / model / ZIP combination is well under $1 per day. The $5/month Apify free credit covers around 1,000 trial listings on the FREE plan.

Frame it as ROI. A flipper finds one $2,000-margin underpriced car per ~200 KBB listings scanned. At GOLD pricing that's about $0.70 in actor cost per $2,000 of margin found, pays for itself on the first flip of the month.

You only pay for records actually saved.

Is it legal to scrape Kelley Blue Book?

Yes for public listing data. KBB's listing pages are served without authentication and indexed by every major search engine. This actor uses standard HTTP requests through Apify residential proxies with realistic browser headers, never bypasses login or paywalls, and never harvests personal data outside the dealer business contact information already published on the listing page.

For commercial redistribution of scraped KBB data (reselling raw inventory feeds), consult your own legal counsel: dealer inventory and KBB's price-advisor data may be subject to KBB's terms of service and your jurisdiction's database-rights regime.

Examples

Example 1: Flipper hunting underpriced 2020-2024 BMW 3 Series within 50 mi of Beverly Hills

{
  "make": "BMW",
  "model": "3 Series",
  "yearMin": 2020,
  "yearMax": 2024,
  "zip": "90210",
  "searchRadius": 50,
  "maxItems": 50
}

Filter the resulting dataset on kbbPriceAdvisor.kbbGrade == 'GREAT' to get the shortlist.

Example 2: Dealer monitoring competitor Toyota Camry pricing under $25K, 100 mi of Dallas

{
  "make": "Toyota",
  "model": "Camry",
  "yearMin": 2019,
  "priceMax": 25000,
  "zip": "75201",
  "searchRadius": 100,
  "maxItems": 100
}

Example 3: Certified pre-owned Lexus RX in the New York metro

{
  "make": "Lexus",
  "model": "RX",
  "condition": "CERTIFIED",
  "zip": "10001",
  "searchRadius": 75,
  "maxItems": 50
}

Example 4: Low-mileage Tesla Model 3 anywhere in California, fleet sourcing

{
  "make": "Tesla",
  "model": "Model 3",
  "mileageMax": 30000,
  "zip": "94103",
  "searchRadius": 200,
  "maxItems": 100
}

Example 5: Ford F-150 trucks all years near Houston, wholesale buyer sweep

{
  "make": "Ford",
  "model": "F-150",
  "zip": "77001",
  "searchRadius": 100,
  "maxItems": 200
}

Example 6: All Honda models within 25 miles of Chicago, broad-make market scan

{
  "make": "Honda",
  "zip": "60601",
  "searchRadius": 25,
  "maxItems": 100
}

Example 7: Insurance underwriter, 2018-2022 Subaru Outback in Denver metro

{
  "make": "Subaru",
  "model": "Outback",
  "yearMin": 2018,
  "yearMax": 2022,
  "zip": "80202",
  "searchRadius": 75,
  "maxItems": 100
}

Input parameters

Field	Type	Default	Description
make	string	required	Vehicle make (e.g., BMW, Toyota, Ford).
model	string	(any)	Vehicle model. Empty = all models of the make.
yearMin	int	(none)	Earliest model year.
yearMax	int	(none)	Latest model year.
priceMax	int	(none)	Cap by asking price (USD).
mileageMax	int	(none)	Cap by odometer reading (miles).
zip	string	required	5-digit US ZIP at the center of the radius.
searchRadius	int	50	Miles. 10-500.
condition	enum	USED	USED, NEW, or CERTIFIED.
maxItems	int	10	Hard cap on saved listings. Raise for production.
proxyConfiguration	object	US residential	Apify proxy settings; US residential is required for KBB.

Kelley Blue Book output format

kbb_listing

Field	Type	Description (and how a buyer uses it)
outputSchemaVersion	string	Date-stamped schema version ('2026-05-09'), pin downstream pipelines safely
recordType	literal	'kbb_listing'
recordId	string	kbb:vin:<VIN> if VIN is structurally valid, else kbb:listing:<id>, drop straight into INSERT ... ON CONFLICT DO UPDATE
listingId	string	KBB's internal listing ID, stable across runs
listingUrl	string	Direct URL to the vehicle detail page on kbb.com
vin	string	17-character raw VIN
vinValid	bool	True when the VIN is structurally valid, filter on vinValid: true to drop garbage rows from your dataset
vinDecoded	object	{ year, manufacturer, country } decoded inline, no NHTSA round-trip
year, make, model, trim, bodyStyle	string/int	Identity fields
exteriorColor, interiorColor	string	Cabin
drivetrain, fuelType, engine, doors	string/int	Mechanical
price, priceUsd	string + numeric	"$31,995" and 31995, use the numeric for sorting and filtering
kbbPriceAdvisor	object	The moat field. { fairPurchaseMin, fairPurchaseMax, kbbGrade }. kbbGrade is GREAT | GOOD | FAIR | HIGH, KBB's own deal-quality grade. Filter for GREAT to surface underpriced inventory.
mileage, mileageMiles	string + numeric	Mileage with units stripped for direct math
listingType	string	USED, NEW, CERTIFIED
daysOnSite	int	KBB's days-since-listing, older = more negotiable
listedAt, scrapedAt	ISO 8601	Real timestamps, not "5 days ago" strings
dealer	object	{ ownerId, name, phone, address, city, state, zip, lat, lon, distanceMi }, dealer geo joined per listing, no re-geocoding needed
images	string[]	Listing photos
description	string	Listing description blurb
isCertified, isNewlyListed, isReducedPrice, isHot	bool	Listing flags
fieldCompletenessScore	int 0-100	Self-filtering signal for downstream pipelines and AI agents
agentMarkdown	string	300-500 char ready-to-paste LLM summary card per record

KBB scraper output example

{
  "outputSchemaVersion": "2026-05-09",
  "recordType": "kbb_listing",
  "recordId": "kbb:vin:3MW39FF05R8E31429",
  "listingId": "778876035",
  "listingUrl": "https://www.kbb.com/cars-for-sale/vehicle/778876035?...",
  "vin": "3MW39FF05R8E31429",
  "vinValid": true,
  "vinDecoded": { "year": 2024, "manufacturer": "BMW", "country": "Mexico" },
  "year": 2024, "make": "BMW", "model": "330i", "trim": "Sedan",
  "bodyStyle": "Sedan",
  "exteriorColor": "Black", "interiorColor": "Black",
  "drivetrain": "RWD", "fuelType": "Gasoline", "engine": "4-Cylinder Turbo",
  "price": "$41,995", "priceUsd": 41995,
  "kbbPriceAdvisor": {
    "fairPurchaseMin": 39800, "fairPurchaseMax": 43200, "kbbGrade": "GOOD"
  },
  "mileage": "12,450", "mileageMiles": 12450,
  "daysOnSite": 18, "listedAt": "2026-04-21",
  "scrapedAt": "2026-05-09T20:00:00.000Z",
  "dealer": {
    "ownerId": 79873, "name": "Pacific BMW",
    "phone": "18889871935", "address": "800 South Brand Boulevard",
    "city": "Glendale", "state": "CA", "zip": "91204",
    "lat": 34.13714, "lon": -118.25501, "distanceMi": 9.34
  },
  "images": ["https://..."],
  "fieldCompletenessScore": 96,
  "agentMarkdown": "🚗 **2024 BMW 330i Sedan**\n💰 $41,995 (KBB: GOOD, fair $39,800-$43,200)\n📏 12,450 mi\n📍 Glendale, CA (9.3 mi away)\n🔢 VIN 3MW39FF05R8E31429\n🔗 https://www.kbb.com/..."
}

During the Actor run

Each run pulls KBB inventory for your make/model/ZIP/radius, normalizes every record, and writes the dataset. Sub-minute typical runtimes.

Two artifacts land in the run's key-value store: OUTPUT (compact run summary with VIN-valid count, KBB Price Advisor count, and error counts) and AGENT_BRIEFING (markdown digest with median price, mileage stats, and the top KBB-GREAT listings ranked by margin). The default dataset is accessible via the Apify dataset API or CSV / JSON / Excel export.

FAQ

How is this different from the free KBB scrapers on the Apify Store?

The two free KBB actors on the Store have single-digit lifetime runs and zero reviews. They leak $ and , straight into the price field, never validate the VIN, drop the dealer's lat / lon, and never surface the KBB Price Advisor grade, the field that actually tells you whether a listing is a deal. This actor parses the canonical KBB inventory payload directly, validates every VIN, returns numeric mirrors of price and mileage, joins dealer geo per listing, and surfaces the KBB grade as a top-level filterable field.

Why does this cost more than the free actors?

If you're feeding the data into a pricing model, dealer-monitoring pipeline, or AI agent, the per-record cost pays back in saved engineering hours within the first week. Free actors break monthly when KBB tweaks a class name, you don't get notified, and your downstream pipeline silently goes empty until you notice. This actor is on a versioned schema with idempotent IDs and continuous maintenance.

Will this break when KBB changes their site?

KBB's underlying inventory payload is far more stable than the rendered HTML, so cosmetic UI changes don't break this actor. When KBB does change the underlying contract, the actor adapts within 24-48 hours and we ship a build update.

How reliable is this for daily production use?

Built for it. Sub-minute runtimes, monitored daily, and maintenance updates ship within 24-48 hours when KBB changes anything that matters.

Can I get the KBB Fair Purchase Price for each listing?

Yes, every record carries a kbbPriceAdvisor object with fairPurchaseMin, fairPurchaseMax, and kbbGrade (GREAT, GOOD, FAIR, HIGH). KBB embeds this in the search-results payload for every listing where it has a band; the field is null when KBB does not.

Why not just use cars.com instead?

Cars.com has more total inventory volume and we ship a companion Cars.com actor you should run alongside this one. KBB is the better source when you need the price-advisor grade (cars.com doesn't have it) and the deepest dealer geo. Most flippers and dealer-intel buyers run both: KBB for the deal-quality signal, Cars.com for the volume and trim breadth.

Can I monitor only listings new since the previous run?

Not in v0.1 as a built-in feature. The listingId and recordId are stable across runs, so you can compute the diff downstream by joining against your previous dataset (INSERT ... ON CONFLICT DO UPDATE). A built-in watchlist mode is on the roadmap for v0.2.

Why is my VIN reported as vinValid: false?

Every VIN we return is checked for structural validity. If a dealer or KBB itself transcribed the VIN wrong (a missing or wrong character), the validity check fails and you can drop the row. The raw VIN is still returned in the vin field; you can re-process it downstream if you want.

Can I integrate with Make / Zapier / n8n / Slack?

Yes. From the actor page, click Integrations and pick your destination. Apify provides native connectors for all of them.

Can I use this with Python?

Yes. Use the Apify Python client: client.actor("skootle/kbb-scraper").call(run_input={...}) and read the dataset.

What's the cheapest way to use this?

The $5/month Apify free credit covers around 1,000 listings on the FREE plan, enough to validate the schema against your pipeline. For production volume, GOLD at $3.50 per 1,000 records is the sweet spot; PLATINUM and DIAMOND drop to $3.00.

Your feedback

Hit a bug or want a feature? Open an issue on the Issues tab rather than the reviews page, and we'll fix it fast (typically within 48 hours).

Why choose Kelley Blue Book Scraper

• Flag underpriced inventory in one query, KBB Price Advisor GREAT / GOOD / FAIR / HIGH deal grade as a top-level filterable field. This is the wedge
• Returns dealer geo so you can radius-filter without re-geocoding, name, phone, address, lat / lon, distance from your search ZIP, joined per listing
• Sub-minute runtimes at low compute cost, daily pipeline finishes fast and stays green
• Every VIN is real. No junk rows polluting your dataset, structurally validated on every record
• VIN decoded inline, model year + manufacturer + country, no NHTSA round-trip, no extra API call
• Sort and filter without parsing strings, numeric mirrors of every money and mileage field
• Hand-tuned for KBB. Fixes ship the same week the source changes, continuous maintenance
• Agent-ready, agentMarkdown per record + AGENT_BRIEFING.md per run, drop into Claude / ChatGPT / Slack as cards
• Schema doesn't break your pipeline, outputSchemaVersion literal on every record
• Safe to dedupe across re-runs, kbb:vin:<VIN> when valid, kbb:listing:<id> otherwise

Other Skootle actors you might want to check

• Cars.com Scraper, companion used-car actor with the same record schema. Run both to combine KBB's deal-quality grade with Cars.com's volume and trim breadth.
• eBay Sold Comps, sold-listing comps with variant-aware grouping for resellers and pricing engines
• SAM.gov + USAspending Federal Contracts, federal contract opportunities with historical award context

Export, run via API, schedule, or integrate with other tools.

Support and contact

File issues on this actor's page (Issues tab), replies within 48 hours. For feature requests, drop them in the same tracker tagged enhancement.