Token-efficient product pagesfor AI agents

Stable schema, already enriched, ready for AI agents. Across all stores.

Parsed
Enriched

API

One endpoint, agent-ready responses

RequestProduct parse endpoint
POST
/v1/products/parsePOST
ResponseExample output
json

Product Representation Comparison

HTML is for humans. JSON is for software. StoreRouter's product format is for agents.

MetricRaw HTMLRendered DOMShopify JSONStoreRouter .md or JSON
Suff.
7
4
5
9
Vivid
7
6
8
10
Search
6
5
6
9
Tokens
544.4 Kprohibitive
823.9 Kprohibitive
3.7 Kcheap
1.2 Kcheap
Parsesevereseveremoderatetrivial
  1. Sufficiency. Are the facts there? What it is, price, sizing, and stock.

    How it's scored

    Are the facts there? A 1–10 judgment of whether an LLM reading this artifact can answer what the product is, what it is made of, fit and sizing, stated price and per-variant buyability, audience, ratings, and what makes it distinctive — scored in the absolute, not relative to other columns.

  2. Vividness. Could you picture it? Images, colour, shape, and material.

    How it's scored

    Could you draw it or pick it out of a lineup? Imagery at usable resolution, colour and pattern specificity, silhouette, scale or measurements, and material texture.

  3. Searchability. Can an agent find it? Names, synonyms, attributes, and filters.

    How it's scored

    Can an agent retrieve this product for varied queries — exact identity, synonyms, descriptive attributes, structured filters, and style, use case or occasion terms?

Why it matters

Built for production agent workflows

Structured data, not HTML

Skip noisy DOM parsing. Get a normalized product object agents can use directly.

Enriched by default

Taxonomy, attributes, variants, and ontology fields are included in one response.

URL in, product out

Use index-only lookup for a predictable, cacheable product view with no live crawl.

Start parsing product pages

Sign up for API access and turn any product URL into agent-ready data.

Sign Up