AI Storefront
Browse the documentation

Seller Scoreboard

Which dealer websites sell the most in your city, at what price, how fast, and who to conquest.

This is the technical reference. For the price, a rendered sample and answers to common questions, see the Seller Scoreboard product page.

At a glance

Price $2.99 per city delivered. Minimum $4.99 per run. A batch is quoted before it runs: the first call returns a total and charges nothing.
Tool name generate_seller_scoreboard
Scope reports:seller-scoreboard
Built for Small Dealer, Dealer Group, Large Retailer

What it does

The seller scoreboard for one to ten US cities: up to 25 seller websites per city ranked by the vehicles that left them over a window of whole months (3 by default), each with its units, its share of the city, the new and used split, the median advertised asking price at exit, the median odometer, the median days to sell anywhere and the median days on the selling site. A city summary (units, the leading site and its share, the top three sites' share) and conquest notes: who dominates, whether a national retailer site outsells the local dealer sites, and the fastest turner. New and used together by default, or one market. The full scoreboards ship as a CSV. From the Sold Summary API: a sale is inferred when a listing leaves a dealer website, every price is the advertised asking price at exit (never a negotiated transaction price); a seller website is the site the vehicle left from, so a group sharing one domain shows as one seller. Use when a dealer, dealer group or retailer asks WHO sells the most in a city, which sites to conquest, how the locals compare with CarMax, Carvana or EchoPark, or how fast the leaders turn. Do NOT use for one dealer's own pricing (use Lot Pricing Scan) or for what to stock (use Trim & Spec Stocking Matrix or Stocking Guide). The sold data has city grain, not ZIP or radius. PRICING: $2.99 per city, minimum $4.99, up to 10 cities per run; cities that cannot be read or have no seller at the floor are not billed. A run of two or more cities requires cost confirmation: call first WITHOUT confirm_cost_usd for an up-to quote (this charges nothing), then again with confirm_cost_usd set to the quoted cap once the user approves.

Arguments

Argument What it is Example
cities Recommended One city per line as "City, ST" (e.g. "Dallas, TX", then "Fort Worth, TX" on the next line), up to 10. Each city is scored and billed separately; a repeated city is scored once. Dallas, TX
city Optional A single city instead of the list (e.g. Austin). Pass state with it. Austin
state Optional The single city's US state, as a 2-letter code (e.g. TX). TX
inventory_type Recommended Which market: all (new and used together, the default, with the split per seller), used, new, or certified pre-owned. all
months Optional Whole months to count units over, ending at the latest data month: 3 by default, 1 to 12. The medians always describe one month. 3
confirm_cost_usd Optional Only for a run of two or more cities (more than the $4.99 floor). After quoting the user the up-to cap and getting approval, call again with this set to the quoted cap. Charged only for cities scored, never above the cap. Leave empty on the first (quote) call. 5.98

What comes back

Asking for it in a chat client

Which dealer websites sell the most used cars in Dallas and Fort Worth?

Score the sellers in Austin, TX for new and used

Running it from a backend

Grant an agent the reports:seller-scoreboard scope, then call the tool over MCP:

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "generate_seller_scoreboard",
    "arguments": {
      "cities": "Dallas, TX\nFort Worth, TX",
      "inventory_type": "all",
      "months": 3
    }
  }
}

See Server-side access for the token exchange and how to read the response.