Four modules, one unified standard
Core is free and open source. Risk, Oversight and Evidence are commercial modules — together they cover the entire EU AI Act flow from signing to audit dossier.
Signing, hashing and verification. The protocol kernel that makes AI output verifiable.
The protocol itself — signs every AI answer so it can be verified mathematically. SDK in Python, JavaScript and Go. Available on GitHub under MIT licence, no lock-in.
Read full specDrift detection and signed alerts. Spots when the AI model starts behaving differently than expected.
Risk Management for AI — monitors output distribution, catches drift and produces signed alerts that flow straight into the compliance log. Hosted by LexiCo or installed on-prem.
Read full specHuman-in-the-loop with signed reviews. Ensures high-risk decisions have documented human control.
Human-in-the-loop portal — reviewers sign decisions with an Ed25519 key. TrustBadge UI (free MIT component) makes the reviewer status visible in the client interface.
Read full specAudit dossiers with ISO 42001 mapping. Generates signed reports ready for regulators and auditors.
Audit dossier generator — builds complete evidence packages for EU AI Act Annex IV systems. Maps TSP receipts to ISO 42001 requirements, signs the entire dossier so it cannot be altered afterwards.
Read full specWant to test one of the modules in your own environment? LexiCo offers pilot programmes for Risk, Oversight and Evidence.