Six surfaces.
One evidence log.
drag / scroll →
Decoy identities
Decoy email identities seeded into fake credential dumps on paste sites. They alert when someone uses one.
james.miller.4f2a91.a91c@… ← decoy
✓ watching for use
Trap facts
Invisible false facts woven into documents. They catch AI knowledge bases leaking your content.
↑ trap fact surfaced · leak confirmed
Prompt-injection firewall
Three layers between users and your model. Adopt it by changing one base URL.
layer 2 ██████░░░ analyzer
layer 3 ███░░░░░░ sanitizer
Decoy assistants
Fake AI endpoints that fingerprint attackers and classify their intent.
Tamper-evident action log
Every agent action sealed with a SHA-256 content hash. Re-verify anytime.
sha256:3c5e91…7e44 · sealed 09:15:20 UTC
Agent runtime protection
Heartbeats, skill scans, and a live action timeline for your OpenClaw agents.
skill scan → BLOCKED · credential harvest
PRICING
Start free. Pay when your traps earn it.
Monthly billing · full comparison →
Free
$0
Kick the tires, catch your first probe.
Shield
$12 /mo
For a production model behind Shield.
Pro
$24 /mo
For teams that need the paper trail.
Enterprise
$120 /mo
For security and compliance owners.