← Back to portfolioIn development

03 / Machine visibility intelligence

Vercatus

An agentic site-intelligence platform that models a website’s content and measures how search and answer engines can discover, understand, retrieve, and cite it.

Portfolio project
StatusIn development
MarketGlobal
ModelB2B SaaS
RoleFounder · Product · Engineering
V

Search visibility now includes answer engines

Companies can rank in traditional search and still be absent from AI-generated answers. Vercatus measures how different machines can reach, interpret, retrieve, and cite a site, then turns the evidence into a roadmap that can be checked in the next run.

A site model built from evidence

A crawl starts from a sitemap or origin, clusters URLs into page templates, and samples them against a page budget. Each page is fetched from multiple perspectives—including browser rendering and named search and AI crawler identities—so the report can distinguish missing content, blocked access, and rendering gaps.

  • 01Template-aware sampling keeps large crawls representative
  • 02Knowledge graph connects entities, claims, intents, and pages
  • 03Local hybrid retrieval simulates answers using the site’s own content
  • 04Specialist analysis is challenged by a critic before findings are published
  • 05Every recommendation is tied to a metric and a measurement cohort

Learn what answer engines cite

Run observations feed a shared repository of page signals, questions, answers, and citation edges. Vercatus compares cited pages within the same question and only surfaces patterns when the evidence clears a support threshold. A Machine Optimization Score and engine-specific prompt tracking help teams follow visibility over time.

  • 01Citation and brand-share tracking across answer engines
  • 02Tracked prompts with answer and ranking changes
  • 03Reverse analysis of pages engines cite
  • 04Fact-checked content rewrites re-scored before presentation
  • 05Durable site memory carries findings and status between runs
TypeScriptEnd-to-end platform
Postgres + pgvectorSite and observation store
Resumable runsIdempotent workflow steps

A working agentic site-intelligence platform

Vercatus is being developed as an integrated research and operating system: CLI and API workflows produce durable reports, while the dashboard organizes visibility, prompts, competitors, intelligence, actions, sites, and settings. Its honest limits are made explicit: crawler impersonation is evidence, not proof of a real block, and simulated retrieval is not the same as live engine behaviour.

Explore another projectPiff↗