The proof
What an agent receives: before vs. after
The same page, twice: once as a browser receives it, once as an agent does.
Commerce PDP capture
98.6% smaller69× less context
Before: raw PDP HTML
91,226 bytes · ~22,789 tokens
After: gateway markdown
1,315 bytes · ~328 tokens
Before: the raw PDP HTML
Theme CSS, mega-menu, product JSON blobs, tracking snippets, a reviews widget, upsells, modals, a cookie banner, a newsletter modal, and cart JS.
Price at character offset 23,185, inside a JSON blob. Stock, shipping and returns sit deeper still.
After: gateway markdown
Price on line 4. Stock, shipping and returns in the first screenful.
# [Alpine Trail Pack 28L](https://basecamp-supply.example/products/alpine-trail-pack-28l) - **Brand:** Basecamp Supply Co - **Price:** $148.00 - **Availability:** In stock - **Shipping:** Free US shipping on orders over $75; standard shipping $6.95. - Free shipping threshold: $75.00 - Ships to: US, CA - Delivery estimate: 3-6 days - **Returns:** 60-day returns. Items must be unused with tags attached. - Return window: 60 days - Policy: https://basecamp-supply.example/policies/returns
Document capture
89.6% smaller
A clinic's services page: 3,388 tokens of themed HTML become 349 of markdown. That is 9.7× fewer.
The opening hours sat at character offset 10,152 of the HTML. In the markdown they are on line 8, with every fee and accepted insurer beside them.
Captured from the services page of a fictional physiotherapy clinic (Fernwood Physiotherapy). No price, no buy button, no product markup, the same measurement on the kind of page most of the web is made of. Two pages, two measurements, quoted separately.
Smaller matters less than the fact that an AI agent can lift these straight into its answer.
The same URL and the same substance in a different format, so browsers get your HTML unchanged and Googlebot always receives the canonical HTML. Captured from a Shopify-style product page for a fictional demo store (Basecamp Supply Co), regenerated and checked automatically.