Kelley Blue Book Scraper

Extract public vehicle listings from Kelley Blue Book cars-for-sale pages into an Apify dataset for automotive research, inventory analysis, and AI-agent workflows.

Use it

Pass a public KBB search URL, for example:

{
  "startUrls": [{ "url": "https://www.kbb.com/cars-for-sale/used" }],
  "maxItems": 25,
  "maxPages": 2
}

The actor collects one record per unique listing detail URL. Fields include the displayed title, year, make, model, trim, advertised price, mileage, price badge, seller text, distance, badges, image URL, source page, and normalized card text.

Pricing

Event	Price
Actor start	$0.00005 per Apify start event (minimum one; memory-based synthetic event)
Vehicle listing extracted	$0.003 per unique listing

With the Actor's 1 GB default, a run capped at 25 listings has a maximum event charge of $0.07505, excluding Apify platform usage. The run status and SUMMARY record report the charged listing-event count and estimated event total.

Proxy configuration

The deployed actor uses encrypted runtime environment variables for the authorized DataImpulse proxy: DATAIMPULSE_PROXY_USERNAME, DATAIMPULSE_PROXY_PASSWORD, DATAIMPULSE_PROXY_HOST, and DATAIMPULSE_PROXY_PORT. Credentials are not part of the input schema or source code. KBB uses Akamai bot protection that blocks datacenter IPs, so a residential proxy is required for reliable extraction; the actor automatically retries on a fresh residential IP when it detects an access-denied challenge. If no DataImpulse credentials are configured, the actor falls back to its configured US Apify proxy, which may be blocked and will report the failure honestly instead of returning empty results.

Scope and limitations

This actor reads publicly accessible KBB cars-for-sale pages. It does not provide KBB valuation data, dealer APIs, VIN-history reports, authenticated information, or a way to bypass access controls. KBB can change its page markup, so selectors are intentionally backed by detail-link discovery and visible-text parsing; review warnings and the SUMMARY record when a run returns no listings.

Start with a small maxItems value. Results reflect advertised listing data at collection time and should be refreshed before making purchasing decisions.

Local development

npm install
npm run build
npm start