FactNotebook Get a proposal
Demo dossiers · synthetic data · shared by direct link

One product. Two modes.
See the shape of the evidence.

Operational Evidence proves whether your AI controls actually operated — from your technical and operational artifacts, no source code. Full Audit adds the code, docs and tests: declared, implemented, observed. Each public summary below is real; the full dossiers are delivered on request.

Runtime behavioural evidence Deterministic · no LLM required SHA-256 sealed EU AI Act Art. 9–15
📊
Operational Evidence

Runtime artifacts only — what the system actually did. No source code. Deterministic, runs air-gapped.

🧬
Full Audit

Everything above + code, documentation and executable tests. Declared vs implemented vs observed.

⚠️
Demo dossiers — synthetic runtime data These audits are published for demonstration; runtime traces were synthetically generated to illustrate the methodology, systems are anonymised. Production dossiers contain live execution evidence and are delivered under NDA — contact@factnotebook.com.

📈 Agentic Trading Platform

Financial AI agents · High-risk · Annex III · Both modes (runtime available)
NON-COMPLIANT
46
Checkpoints
57.2/100
Compliance score
14
Reports
SHA-256
Sealed

Full four-phase audit: static analysis, executable tests, runtime behavioural evidence and a governance status board. Runtime traces show iteration overruns and hallucinations. Sincerity gap: 20.5%.

📄 Executive Summary (public)
📊 Operational Evidence runtime · no source code
🔒
Behavioral Audit
Runtime evidence, system health
🔒
Decision Accountability
Named decision-makers ↔ runtime
🔒
Risk Control Matrix
Business risk exposure
🔒
Runtime Trend
Compliance over time
🔒
Runtime Sessions
Session-level behavioural logs
🔒
Governance Board
Radar, checkpoint matrix
🧬 Full Audit + code · docs · tests
🔒
Executive Report
Score, verdict, key findings
🔒
Sincerity Report
Declared ↔ observed gap, 20.5%
🔒
Technical Report
Executable tests, article-by-article
🔒
System DNA
Architecture, component map
🔒
Evidence Ledger
Full evidence chain
🔒
Remediation Playbook
Prioritised fixes, score projection

🏥 MedAI Library

Healthcare AI · High-risk · Annex III · Full Audit only (no runtime connectors)
PARTIAL COMPLIANCE
46
Checkpoints
Art. 9–15
Coverage
8
Reports
SHA-256
Sealed

Static analysis and executable tests. No runtime connectors were wired — this dossier illustrates a Full Audit without operational traces (declared and implemented, not yet observed at runtime).

📄 Executive Summary (public)
🧬 Full Audit + code · docs · tests
🔒
Executive Report
Score, verdict, key findings
🔒
Sincerity Report
Declared ↔ code contradiction
🔒
Technical Report
Executable tests, article-by-article
🔒
System DNA
Architecture, component map
🔒
Evidence Ledger
Full evidence chain

Want a dossier on your system?

Production audits with live runtime traces, re-executable tests and SHA-256 sealed evidence — delivered under NDA. Operational Evidence can run air-gapped, with no source code and no LLM.