This site's own operational layer
Share with Agent doesn't just sell the operational layer — this site runs one, and it's the living demo. Here is everything we publish about ourselves, machine-readable and live:
The endpoints
| What | Where |
|---|---|
| Machine twin of every page | <path>.json — e.g. /index.json, /packs.json, /blog/….json |
| The hub (all twins) | /agent.json |
| Convention declaration | /llms.txt |
| Classic siblings | /robots.txt · /sitemap.xml |
What's inside each twin
- Transparency declaration — who publishes this, and that it's open vendor guidance for the user's agent, never hidden.
- Consent gate — the agent asks its user; a yes authorizes, one-time or per operation.
- User-Intent Matcher — declared audiences plus limited gap questions used only when known intent is insufficient.
- Adoption + Operation guidance with the Service Control Map — how each operation (evaluate · enroll · transact · configure · use · manage) is actuated: API endpoints, functions/actions, interface controls.
- Live data — pricing twins are generated from the database on each request from the current pricing database, subject to normal caching and service availability.
The share codes
Every page's share control copies a code like this:
swa:1:<base64url-of-the-manifest-url>The scheme is swa, version 1, then the page's manifest URL,
base64url-encoded. Any AI can decode it and fetch the manifest. (Reference encoding is
an optional embodiment — QR codes and plain URLs work the same way.)
The discovery pointers
Agents find the layer through multiple signals on product pages and articles: a
<link rel="alternate" type="application/json"> in the head, an HTTP
Link header, the visible footer line, and llms.txt. Terms,
Privacy, Refund and Acceptable Use remain human-only legal pages and do not carry
machine-directed guidance or machine twins.
Want this exact setup on your site? Start with one hosted page free, or join the waiting list for more pages or optimization.