Five surfaces. One archive.

A deep read of what is teaching AI about you. Sweep renders your domain, your category’s citation graph, and the pages agents actually consult, the way agents receive them, not the way a browser shows them.

• Dual-view reads: human render vs. crawler/agent response
• Hidden-instruction detection across headers, comments, metadata, CSS-hidden blocks, Unicode smuggling
• Planted-narrative signals: seeded claims, review clusters, citation anomalies
• Output: one verdict and one Evidence Pack, not a score you explain away

Sweep is a photograph. Patrol is the film. Once the archive knows your baseline, every new hidden instruction, planted page, or citation-graph shift is a diff, not a discovery problem.

• Baseline + delta monitoring across your source graph
• Campaign clustering: one planted page is noise; forty domains moving together is a campaign
• Confidence-scored alerts that attach the Evidence Pack inline

The product is not a dashboard. It is a document that survives scrutiny: what was found, where, when, how it reads to an agent, what it propagates into, and the shortest path to remove it.

• Source capture + timestamp + crawler/agent view
• Attribution trail and propagation analysis
• Takedown playbook: platform report, hosting-abuse route, legal posture notes
• Formatted for counsel, crisis-PR, and platform trust & safety teams

Source-integrity signals as a feed for GEO platforms, digital-risk-protection vendors, insurers, and legal teams who need the data layer without rebuilding the archive themselves.

• Campaign and infrastructure intelligence across the agent-readable web
• API and licensed feeds; white-label research
• The strategic position: the layer others integrate

White-label source integrity for the agencies who already own the client relationship. Your cover, your engagement, our archive and verdicts underneath.

• Built for GEO/AEO agencies, crisis-PR firms, and brand-protection consultancies
• Client-facing reports under your brand
• Now recruiting design partners. bring us your hardest case

Two honest tools that show the second web exists before anyone pays for anything:

• cloak check: does your page look different to a human, a search crawler, and an AI crawler? Cloaking detection as a public good.
• answer trace: which sources are feeding the AI answers about you right now? The citation graph behind the answer, as evidence.