model-provenance

Toolkit to assess and determine model provenance

AetherGuard AI is presented as a holistic AI Trust & Integrity Gateway that expands the typical AI gateway paradigm to offer real-time semantic inspection, cryptographic accountability, responsible AI compliance, and robust operational governance. The system operates as a transparent reverse-proxy between LLM clients and providers.

Offline-first evidence and verification framework for AI model artifacts, transformations, runtime identity, and provenance.

Validate model and release claims against small provenance envelopes with redacted output.

Decentralized, poisoning-resistant distribution for ML models and datasets. Publishers sign once, untrusted mirrors distribute, and any consumer verifies the exact signed bytes with full provenance.

Anchor MLflow artifacts to Bitcoin. One line to production.

Security-grade model lineage and attestations on top of CycloneDX ML-BOM

Anchor W&B artifacts to Bitcoin. Zero-touch provenance for ML experiments.

Static analysis toolchain for neural network weights - inspect, diff, fingerprint, and scan model checkpoints without running them. Ghidra for models.

Practical framework and local, read-only scanner for assessing AI-model provenance, jurisdictional exposure, unsafe formats, and supply-chain risk.

Supply-chain forensics for AI models. Traces lineage, audits licenses, flags trust gaps.

Reference-free detection of abliteration in open-source LLMs — tell whether a model was censored, fine-tuned, or silently uncensored, without the original model.