Send us the same inventory feed your other vendors already get. We turn every vehicle into pages, a live endpoint, and outbound listings designed for AI discovery, so assistants can read your inventory and link inquiries back to you. Live in about an hour.
Find me a reliable AWD SUV under $18k near Hamilton, good on gas.
A strong match is the 2019 Subaru Forester 2.5i at Maple Ridge Auto. AWD, 74,200 km, about 8% below the regional median.
$17,900all-in
The same feed you already hand your other vendors. You point us at your existing export once, and you are done. No code, no new website, nothing to install.
Use what you already have
Most dealers already produce a feed for their website or marketplace vendors. We take that same file. There is nothing new to build.
Wired up once
You give us secure access to the feed a single time. After that it refreshes on its own, on a one-hour freshness target.
Nothing scraped
We only read the feed you authorize. We never crawl or copy your website to pull inventory.
We pull over SFTP (secure file transfer, pull preferred, so there is no push infrastructure to stand up on your side). Each provider feed runs through a dedicated feed-format adapter, one parser per source (EDealer, D2C, Convertus, Dealerpull, and so on), that maps the file into our single canonical vehicle schema. These are format adapters, not scrapers: we read the structured feed you authorize, and only that. HTTP delivery is supported where a provider offers it.
Sample pipeline state, illustrative. A pilot feed maps from the dealer’s export to canonical.
On financing
Any financing example shown is illustrative only, a sample calculation to help a shopper picture a payment. It is never an advertised rate, term, or approval. Vehicle prices are shown all-in, the price a shopper actually pays, with no hidden mandatory fees.
We check each vehicle against official records and live market data, then write a clean, factual summary. The listing reads accurately to a buyer and to an AI assistant.
Specs checked
Every VIN is validated against the official US and Canada vehicle database, and gaps in your feed are filled from a decode source.
Market context added
Factual, informational context, for example "priced 8% below the regional median for this year, make, and model." Context only, never a pushy claim.
Clean condition summary
One tidy, accurate description per vehicle. No invented features, no templated filler.
Validation runs against NHTSA vPIC (the free US federal VIN decoder) for spec correctness, with MarketCheck NeoVIN used to decode and gap-fill missing fields, cached so we never re-decode the same VIN needlessly. Market context comes from MarketCheck’s market dataset. The condition summary and any financing scenario come from a single prompt-cached enrichment call per vehicle, keyed so it only re-runs when the description, options, or photos actually change. A price or mileage tweak never triggers a re-write. Financing math is rendered as a clearly labelled illustrative scenario with disclaimers, not an offer.
Every validated vehicle goes live across four surfaces at once. Two of them work by being read when a buyer is already looking, your listing pages and the live endpoint. A third pushes your inventory into the AI shopping catalogs ahead of time, so your cars are in the index before anyone searches.
Fast, static, schema-rich, and branded to you. Crawlable and structured, so crawlers can read the whole vehicle without running a line of JavaScript.
2019 Subaru Forester 2.5i
JF2SKAGC9KH, 74,200 km
$17,900 all-in
Assistants do not just read the page. They can ask your inventory a direct question and get a straight, attributed answer back. Walk through it just below.
AI-referred visits, agent tool calls, and the leads they produce. Every figure is a confirmed event, never a made-up reach number.
Your inventory formatted for each AI shopping catalog and refreshed automatically, published to each as the channel opens to feeds.
Most listings are a flat page an assistant has to read top to bottom. This is different: the assistant can ask a direct question and get a precise, attributed answer back, pulled live from the feed you authorized.
What the buyer sees
Find me a reliable AWD SUV under $18k near Hamilton, good on gas.
A strong match is the 2019 Subaru Forester 2.5i at Maple Ridge Auto. AWD, 74,200 km, about 8% below the regional median.
$17,900all-in
What the assistant receives
Sample record, illustrative. The technical name is MCP, an open standard AI tools use to pull live data. Every answer links and attributes back to you.
Surfaces 1 and 2 wait to be read. This one does not. We format your inventory the way Microsoft, OpenAI, and Google each want it and feed it straight into their shopping catalogs, so when a shopper browses cars inside an assistant, yours is already in the index to be shown.
What the buyer does
“Show me AWD SUVs under $18,000 near Hamilton.”
asked inside Copilot shopping
The assistant answers from its shopping catalog, the index we feed your inventory into, not from a live crawl of your site.
Where your car shows up
Catalog formats: UCP JSON for Microsoft, JSONL for OpenAI, a merchant feed for Google. Refreshed about every four hours.
One honest note. These shopping catalogs are shared by every seller, the way they are built, so this is the one surface where your car can appear near another dealer’s. Even there, your result carries your own link and attribution and the inquiry still routes to you, while your own listing pages and endpoint stay completely competitor-free. Several of these programs are still rolling out, so we publish to each the moment it opens to feeds.
Listing pages are statically rendered with Vehicle JSON-LD in the initial HTML response, which matters because most AI crawlers cannot execute JavaScript. The agent endpoint is a server MCP route (Model Context Protocol over HTTP) plus browser WebMCP tool declarations, with read tools like search_inventory and get_vehicle. The dashboard is powered by first-party server-side analytics on two axes, hit type by agent platform, with Web Bot Auth verification for signed agents. Outbound syndication writes UCP JSON and OpenAI JSONL feed files on a roughly four-hour refresh.
A shopper asks an AI assistant for a car. When yours is surfaced, the links and attribution route the inquiry straight back to your listing, and the lead lands in your CRM with the AI source attached.
Eligible for citation, ready for the click
When the assistant cites your listing, it links straight to it, and the buyer lands on a page that is unmistakably yours.
Lead goes to your CRM
When they reach out, the lead posts to your CRM with the AI source captured, so you know where it came from.
Always your customer
The buyer contacts you and buys from you. We make your vehicle the cited source, not one listing in a competitor feed.
Each listing carries a dealer-branded lead form that posts to your CRM with full agent-attribution metadata, delivered as ADF/XML (the format most dealer CRMs ingest) or a JSON webhook. Confirmed AI-referred visits and agent tool calls are recorded as first-party events you can see in the dashboard. There is no cross-dealer carousel, no "similar vehicles from other dealers," no competitor carousel and no other dealers’ cars surrounding yours. The boundary is structural, not a promise.
Our promise
We never scrape your inventory, and your own listing pages and endpoint never sit your vehicle next to a competitor.
Your inventory comes only from the feed you give us. A marketplace places your vehicle alongside competing dealers on its own site and recommends other cars. The AI shopping catalogs we push to are shared by every seller, but even there your car keeps your link and attribution and the lead comes to you. On the surfaces that are yours, the page and the endpoint, there are no competitors at all. We make you the cited source. That is the whole model.
AI platforms decide what to crawl, cite, and rank. VIN Index gives your inventory the strongest technical foundation to be discovered.
Run the free check on your own site and see exactly how visible your inventory is to AI right now. No feed, no setup, no card.