Machine-Readable Artifact

rtfct: The Attestation

What It Is

The native cryptographic record format. Not a dashboard. Not a feature. It is the proof.

Overview

Every AI interaction leaves a mark. rtfct captures the complete state — input payload, output response, model version, policy evaluation matrix, precise UTC timestamp, and SHA3-256 cryptographic hash — in a unified format that proves its own structural integrity. It cannot be altered, retrofitted, or deleted without invalidating the entire hash chain. It is the native language of AI accountability. When an enterprise must present its operational state to a court, regulator, or internal audit committee, rtfct exports as The Attestation — the court-ready presentation layer.

Technical Specification & Key Features

Self-Verifying Structure: Embedded cryptographic integrity proofs built directly into the JSON-LD schema, eliminating third-party validation dependencies.
Tamper-Evident Hash Chain: Any alteration to a single bit of historical telemetry breaks the mathematical verification chain, alerting compliance sentries immediately.
Cross-Jurisdictional Admissibility: Structured to satisfy the stringent Federal Rules of Evidence (Rule 902(13)/(14) for self-authenticating electronic records), the EU AI Act Article 12 record-keeping mandates, and UK evidence frameworks.
Immutable Append-Only Filing: Once an rtfct unit is compiled, it enters an append-only state with zero post-generation editing or deletion vectors.
The Attestation Export: Single-click compilation into a court-ready, human-readable presentation package featuring embedded cryptographic verification seals.

Evidentiary Use Cases

Litigation Defense: Produce the exact, unalterable record of an AI inference decision during formal discovery proceedings to refute claims of algorithmic negligence.
Statutory Regulatory Audit: Prove continuous, active compliance under the EU AI Act, Colorado SB 24-205, California AB 2013, and FTC enforcement mandates.
Board & Institutional Governance: Document model lineage, system decisions, and guardrail performance for permanent corporate memory across model migrations.