Capabilities
Every agent action. Every decision. Every output.
Recorded, timestamped, and hash-chained in real time. Immutable by design. Court-admissible by construction.
What the Trace records
The Trace is not a log. A log is append-only by convention. The Trace is hash-chained — each entry references the hash of the entry before it. Modifying any entry breaks every subsequent hash. Tampering is mathematically detectable.
Every Trace entry records:
Hash chain
Breaking any entry in the chain breaks all subsequent entries. This is not encryption. It is structural integrity — the same principle used in blockchain and certificate transparency logs. A modified Trace is a detectable Trace.
Compliance
The EU AI Act requires high-risk AI systems to maintain logs sufficient to enable post-hoc auditability. Article 12 specifies: automatic logging of events throughout the AI system's lifetime, retention periods aligned with intended purpose, and records sufficient to assess compliance.
The Trace satisfies Article 12 by design.
Every TraceEntry includes the euAiActRiskLevel and euAiActLogVersion fields. The retention policy is configurable per engagement. The hash chain provides the auditability the regulation requires.
Evidence chain
When a dispute is filed, Parler reads the Trace. The panel evaluation receives:
—The compiled Paper (what was agreed)
—The complete Trace (what actually happened)
—The behavioral scores at each step
—Any APEX-BG enforcement events
The hash chain makes the Trace tamper-evident. Neither party can modify the record after the fact. The evidence is symmetric — both Buyer and AI Provider see the same Trace.