One readiness score. Separate technical evidence.
Understand what AI can find, read, and use on your website. Four readiness pillars contain 15 scored checks. Fourteen separate technical checks provide supporting evidence.
See what we measure- 01Start with your public website
No login needed. No forms submitted.
- 02Get your readiness picture
Four pillars. 15 scored checks. One score out of 100.
- 03Look closer at the evidence
14 checks, with clear findings and next steps.
Four pillars. A clearer picture.
Each pillar contributes up to 25 points, for a total of 100. Your result link keeps the same scan available to revisit or share.
Extraction
Can the scanner identify useful business facts and content from public pages?
Discovery
Can important business information and machine-readable entry points be found?
Agentic Tool Execution
Can agents discover server-side business tools and supported actions through MCP, tool manifests, and Action Cards?
- MCP endpoint: 13 points
- Tool manifests: 8 points
- Action Cards: 4 points
Bot accessibility
Can automated systems reach the public site and interpret its access signals?
Technical evidence explains the site. It does not change the score.
Each check shows what passed, what needs attention, or what could not be checked. These findings have their own technical score, kept separate from your four-pillar readiness score.
Access & transport
DNS and HTTPS, AI-relevant headers, and robots.txt policy.
Discovery
llms.txt, sitemaps, well-known files, API discovery, and feed freshness.
Business context
Structured data, local identity, pricing transparency, and social metadata.
Content quality
Page weight and whether public content is practical to quote and reuse.
How to interpret a result
The action plan orders failed and warning checkpoints by the points actually deducted by the service. It shows available points and returned remediation text, without guessing effort or promising a future score.
What the scanner does not measure
The result does not predict search ranking, AI citations, conversion, security, compliance, or whether a third-party model will be correct. It does not access credentials or submit forms.
History and technical summaries stay on this device.
Tracy keeps at most 20 recent result references in browser storage for history and comparison. It does not store fetched page bodies or raw website content there.
Exact shared links restore the persisted four-pillar result and explanation. Live tool evidence, optional profile checks, and single-tool rechecks remain on the device that ran them.
Settings control advanced rechecks and the preferred evidence view. Website captures stay only in the active browser session, while the AI narrative is cached by the scanner service.