SchemaX Docs

SchemaX keeps your site legible to machines, continuously, proves it, and shows you what it earns you. There is one verified graph of your business, and everything below is compiled from it.

Build and verify the graph:

Discover: builds a machine-readable business profile from your site and keeps it in sync as your content changes. This is what makes ChatGPT, Perplexity, and Google AI Overviews understand and recommend you.
VMP (Verified Machine Profile): an optional trust layer that gives your profile a verifiable status signal for both humans and machines.

The acting layer turns the verified graph into agent-facing surfaces, and measures what they earn:

Your site as an agent app: a hosted MCP server, one per site, that lets agent users ask about your business and send an inquiry, answered only from verified content.
Outcome measurement: a small snippet that shows AI-referred visits, agent activity, crawler hits, and the conversions that follow.
Agent-commerce readiness: the /.well-known/ucp manifest and the ACP product feed with its gap report, built to never claim more than is true.

Each acting-layer surface is compiled from the same verified graph and is off until you turn it on. A human approves anything that goes live.

Note on Convert. The former Convert experiment workflow has been retired. Its measurement layer lives on: “Convert” now means turning the verified graph into agent-facing surfaces and agent traffic into measured outcomes. See Outcome measurement and Your site as an agent app.

Where to start

If you’re new, go in this order:

Getting started: create your first project and run a scan
Discover guide: review, release, and deploy your machine-readable profile
VMP: add a verified status layer once your profile is clean

Then add the acting layer:

Your site as an agent app: the per-site MCP server
Outcome measurement: measure what machine legibility earns
Agent-commerce readiness: the UCP manifest and ACP feed
WordPress plugin: deliver everything to a WordPress site in one install