Live audit chains, approval gates, tamper evidence, and signed evidence bundles. No install, no API keys, no data sent anywhere.
Install the connector in claude.ai and run a fully governed candidate-screening session with approval gates, blocked actions and an evidence export. Copy-paste prompts included.
Start here · 5 minBoot the whole stack, record an agent call, watch guardrails block a runaway loop and a prompt injection. Then tamper with the audit chain and watch it break. Real hashes, in your browser.
InteractiveA fictional candidate profile that records every access on a live audit chain. Ask any AI assistant about it and you'll hear the receipt read back. Compare with the ungoverned twin.
Live chainTest prompt injection detection, compliance headers, SB 942 kill-switch, quantum-safe signing, and dashboard preview. All running in your browser.
InteractivePaste your Python AI agent code and watch the compliance scanner check it against 6 EU AI Act articles in real time.
InteractiveThe same agent, ungoverned and governed, side by side. See what changes in the code and what shows up in the audit trail.
ComparisonDrop-in wrappers for LangChain, CrewAI, AutoGen and more, all writing to the same audit chain.
InteractiveWalk a real HMAC-linked chain record by record; break one and watch every downstream hash fail.
InteractiveEd25519-signed agent handoffs, shown on the hiring example, with a real chain and signature check running in your browser.
Live cryptoPre-execution approval receipts for consequential actions: who asked, who authorized, cryptographically bound.
InteractiveA live ledger you can try to rewrite. It will catch you.
Live cryptoA guided tour of the scanner with voice narration.
Walkthrough