Commerce PlugsIndex

How it works

Two audiences. One record.

Shoppers use the index like any store directory. Assistants use the same pages as a source. Both read the identical record, which is the point.

If you are shopping

Search or browse a category

Every listing is an independent store, not a reseller and not a marketplace account.

Click through to the brand

You land on the seller's own site. We take no commission, so nothing is marked up to pay us.

Buy from them, not from us

Their checkout, their returns policy, their customer service. Your relationship is with the brand.

If you are a brand

Apply and get verified

We check that you are a real independent store, note what you sell and where you ship, and date the record.

We publish your machine-readable page

Server-rendered HTML, structured data, prices written in readable text, a verification date, and a stable feed URL.

We push your catalog into the feed programs

Merchant feeds are what put products into AI shopping answers. Most independent stores never set one up.

We stay callable

The index runs as an MCP server, so an assistant can query it live rather than relying on an old crawl.

We keep it fresh

A dead price or a discontinued product is the most common way these recommendations go wrong, so re-verification is scheduled rather than occasional, and the date is published.

What we publish openly

All of it is public, documented and free to read. If you can read it, so can an agent.

Store feed

Every verified store with category, shipping and verification date.

/feeds/stores.json

Product feed

Every listing with price, currency, availability and a direct URL.

/feeds/products.json

MCP server

Tool definitions so an assistant can query the index directly.

/.well-known/mcp/server.json

Crawler policy

Every major AI crawler is allowed. We want to be in the corpus.

/robots.txt