TauDIL Platform

Deterministic
Intelligence
Layer

The governance platform that guarantees safe, explainable, and compliant AI decisions. Every verdict deterministic. Every decision auditable. Every action cryptographically bound.

TauDIL PRO · Live
2,153
Rules Executed
1,268
Authentications
Enforcement Actions843
Events Recorded9,772
Integrity Score99.98%
Avg Latency<500ms
System ModeOPERATIONAL
Rules Executed 2,153 ▲15.3%· UAE Engine ACTIVE· EU AI Act 94.4%· Aelthered Chronicles RECORDING· SYGON Coherence ENFORCING· TRCP-Φκ Scoring LIVE· AGL Admission ACTIVE· Zero Violations · Rules Executed 2,153 ▲15.3%· UAE Engine ACTIVE· EU AI Act 94.4%· Aelthered Chronicles RECORDING· SYGON Coherence ENFORCING· TRCP-Φκ Scoring LIVE· AGL Admission ACTIVE· Zero Violations ·

Deterministic Rules Creation

TauDIL operates two distinct rule systems: Assessment Rules via the UAE Engine for business decisions, and AI Governance Rules via the USE Engine for all AI interactions. Same input always produces the same output.

Operator
Syntax
Example
>
field > value
amount > 50000
<
field < value
credit_score < 600
>=
field >= value
loan_to_value >= 0.85
==
field == value
status == "sanctioned"
IN
field IN [v1, v2]
risk_rating IN ["HIGH", "CRITICAL"]
CONTAINS
field CONTAINS val
description CONTAINS "override"
MATCHES
field MATCHES regex()
input MATCHES regex(jailbreak)
IS NULL
field IS NULL
audit_file IS NULL
AND / OR / NOT
cond1 AND cond2
amount > 10000 AND risk == "HIGH"
APPROVE

Decision approved. Logged to Aelthered Chronicles. No human action required.

REVIEW

Escalated for human review via authority matrix. Creates SLA-bound escalation record.

BLOCK

Hard stop. Logged with full evidence chain. Cannot be bypassed by any means.

FLAG

Non-blocking. Decision proceeds but compliance officer notified and logged to Mirror.

Priority
Rule Type
Examples
95–99
Security & Hard Blocks
Sanctions, prompt injection, PII leakage, fraud
85–94
Regulatory Compliance
KYC failures, AML thresholds, Basel limits, GDPR
70–84
Business Rules
Authority matrix, approval workflows, underwriting
50–69
Advisory Flags
Low coherence warnings, boundary nudges, reviews
0–49
Informational
Logging rules, audit markers, notifications
Field
Description
input
Raw user prompt / message sent to AI
output
AI model's response (evaluated post-generation)
kappa
TRCP-Φκ coherence score of AI output (0.0–1.0)
authority_level
User's authority level from profile (0–10)
domain
Domain context: credit-risk, aml-monitoring, etc.
intent
Detected intent from USE classification layer
Priority 99 · Compliance
Applicant On Sanctions List
AML · 6AMLD · FATF
sanctions_status == "MATCH" OR sanctions_status == "PROBABLE_MATCH"
BLOCK
EU AML Directive 6AMLD compliance — mandatory hard block on sanctions matches. Reference GDPR Art. 10.
Priority 98 · Compliance
High-Risk Jurisdiction Block
Trade Finance · FATF · 6AMLD
destination_country IN ["IR","KP","SY","RU","BY"] OR origin_country IN ["IR","KP","SY","RU","BY"]
BLOCK
FATF grey/black list — Article 18 enhanced due diligence. Requires Compliance Director sign-off.
Priority 99 · Security
Prompt Injection Guard
AI Security · USE Layer 1
input MATCHES regex(ignore.previous|system.prompt|forget.instructions|DAN|jailbreak)
BLOCK
USE Layer 1 prompt injection defence. Session flagged and logged to Aelthered Mirror.
Priority 90 · Compliance
Low Kappa Coherence Block
EU AI Act Art.9 · TRCP-Φκ
kappa < 0.45 AND domain IN ["credit-risk","insurance-underwriting","aml-monitoring"]
BLOCK
Minimum coherence threshold for high-risk AI domains. Response withheld. Human review required.
Priority 90 · Fraud
Potential Transaction Structuring
AML · FATF Recommendation 20
transaction_count_24h > 5 AND transaction_amount < 10000 AND transaction_amount > 7000
REVIEW
Multiple transactions below reporting threshold. Potential structuring. SAR generation required.
Priority 88 · Compliance
Investment Advisory Boundary Guard
MiFID II · Article 24
output CONTAINS "you should buy" OR output MATCHES regex(invest.*now|buy.*immediately)
FLAG
AI must not provide personalised investment recommendations without suitability assessment.

Structural Governance
in Autonomous Systems

Security shifts from perimeter firewalls to embedded internal governance. Static boundaries leave internal execution blind. TauDIL's mesh topology replaces boundary firewalls with validation gates between every module — continuous internal routing with sub-millisecond constitutional validation.

S-01

Dynamic Agent Identity

Cryptographic Attestation

Static credentials replaced by cryptographic handshakes validating agent context instantly. The Cryptographic Attestation Node intercepts every agent request before it reaches the Policy Evaluation Engine.

Attestation Latency<1ms
MethodSub-millisecond
Channel TypeAuthenticated Only
S-02

Constitutional Runtime Validation

Merkle-Proofed Baselines

Policies fail under stress. Merkle-proofed constitutional baselines prevent drift. Kernel policy evaluates state transitions pre-execution — not at the I/O boundary. Wrappers cause 120ms latency and bypass risk. Continuous runtime enforces at <1ms with full state scope.

Middleware Latency120ms
Continuous Runtime<1ms
Bypass RiskNone
S-03

Hard Operational Boundaries

Layered Isolation Architecture

Internal layers enforce strict resource and network limits to auto-isolate overruns. Sandboxing Layer → Network Egress Filter → API Rate Limiter → External Boundary. Core AI Runtime operates inside bounded perimeters enforced at every layer.

Architecture4-Layer Isolation
EgressFiltered
Auto-IsolationOn Overrun
S-04

Immutable Audit Chain

ED25519 + SHA-256 Chained

Event batches hashed and chained to create tamper-evident lineage for audits. Event Trigger → Metadata Capture → Hash Generation → Link to Previous Block → Append-Only Ledger. Every interaction feeds a gapless ledger automatically.

Hash AlgorithmSHA-256 Chained
SigningED25519
Tamper-EvidentYes
S-05

Executable Infrastructure

Policy-as-Code · Runtime Compilation

Static policy fails at runtime — manual rules lag behind agent velocity. TauDIL compiles compliance logic natively into system layers. Policy-as-code transforms mandates into deterministic constraints inherent to operations. Governance compiles into architecture.

Static Policy Coverage65%
Continuous Validation98.7%
Audit Trail Completeness99.9%
S-06

Closed-Loop Validation

Continuous Boundary Enforcement

Feedback loops embed boundary checks into execution pipelines. AI Agent → Validation Gate → Boundary Check → Integrity Scan → Execution Approval → back to Agent. Continuous checks maintain strict limits while producing immutable audit trails.

Loop Stages5 Gates
ScopeFull State
Gaps EliminatedYes

AGL™ — Semantic
Admission Controller

Every entity extracted from a document must pass through AGL before it can be added to the Company Knowledge Graph (CKG). AGL uses SYGON's wave-based coherence and drift signals to make a four-way admission decision — preventing contamination, hallucination, and semantic drift from anchoring in the φ-lattice.

ALLOW
×1.0

Entity coherent and grounded. Added to CKG at full confidence.

RESTRICT
×0.50–0.70

Admitted with confidence penalty. Below coherence threshold or high drift.

DEFER
×0.0

Cannot be admitted yet. Staged for human review via admit_deferred().

BLOCK
×0.0

Entity rejected. Contradiction in CKG or confidence < 0.30. Audit log only.

Domain
coh_min
drift_max
Use Case
operational
0.55
0.60
Production systems, live decision support
governance
0.55
0.55
Regulatory compliance, AI Act — strict + conflict always blocks
default
0.45
0.70
General purpose domains
research
0.35
0.80
R&D, exploration, permissive ingestion
Condition
Decision
Mult.
Contradiction in CKG detected
BLOCK
0.0
Entity confidence < 0.30
BLOCK
0.0
No context match AND confidence < 0.65
DEFER
0.0
No context match AND confidence ≥ 0.65 (bootstrap)
RESTRICT
0.50
Context match AND coherence < coh_min × 0.6
DEFER
0.0
Context match AND coherence < coh_min
RESTRICT
0.60
Context match AND drift > drift_max
RESTRICT
0.70
Context match AND coherence ≥ coh_min AND drift ≤ drift_max
ALLOW
1.0
Signal
Type
Description
coherence
float 0–1
Semantic stability of token in wave context. High = well-grounded.
drift
float 0–π
Angular drift in golden spiral lattice. High = semantically unstable.
conflict
bool
CKG contradiction. Async via /conflicts endpoint in hot path.
context_match
bool
Token's dominant wave context is non-general domain frame.

Live Compliance Scores

TauDIL continuously enforces compliance controls at runtime. The following scores reflect real-time enforcement state across all active regulatory frameworks — not periodic audit results.

Review Needed
GDPR
General Data Protection Regulation
76.8%
5 passed · 1 failed · 3 partial
Review Needed
HIPAA
Health Insurance Portability Act
85.3%
7 passed · 0 failed · 3 partial
Review Needed
SOC 2 Type II
Service Organization Control 2
80.6%
7 passed · 0 failed · 5 partial
Review Needed
ISO 27001
Information Security Management
81.3%
6 passed · 0 failed · 4 partial
Compliant
EU AI Act
EU Artificial Intelligence Act
94.4%
10 passed · 0 failed · 1 partial
[ISO 27001] A.9.2.3
Management of Privileged Access
Ensure both admin and user roles exist and are distinct. Score: 50%
HIGH
[HIPAA] §164.308(a)(1)
Risk Analysis
Risk analysis rules not configured. Score: 40%
HIGH
[GDPR] Art. 33
Breach Notification Readiness
Configure security rules for automated incident detection. Score: 50%
HIGH
[SOC 2] CC6.3
Role-Based Access Control
Ensure both admin and user roles exist and are distinct. Score: 50%
HIGH
[GDPR] Art. 5(1)(e)
Retention Policy
Check error: name 'RETENTION_DAYS' is not defined. Score: 0%
MEDIUM
[EU AI Act] Art. 62
Incident Reporting Readiness
Configure security rules for automated incident detection. Score: 50%
MEDIUM

TauDIL maps EU AI Act articles directly to executable enforcement controls. Each article is backed by a structural mechanism — not documentation.

Article 12
Record-keeping
→ Aelthered Chronicles
Article 14
Human Oversight
→ UAE Escalation Engine
Article 15
Accuracy
→ TRCP-Φκ Scoring
Article 27
FRIA
→ FRIA Gate

Aelthered Chronicles &
Runtime Event Lineage

A cryptographically signed immutable operational lineage framework that preserves verifiable execution history across runtime activity. Governance is treated as executable runtime infrastructure — not static policy documentation.

1,268
Authentications
▲12.4% vs 24h
843
Enforcement Actions
▲8.7% vs 24h
2,153
Rules Executed
▲15.3% vs 24h
9,772
Events Recorded
▲18.6% vs 24h
99.98%
Structural Integrity
✓ Intact
Authentication Activity
Enforcement Continuity
Rules-Engine Execution
Chronicle Integrity
Session Governance
Runtime Event Lineage
Escalation Conditions
Execution Accountability
Event ID: 8f3c7a2e-9b1d-4a7f-8c2e-2b6f5e3d9a11 Timestamp: 2025-05-18T14:32:11.Z Actor: system.user@enterprise.com Action: Policy Enforcement Executed Rule Set: Access.Control.v2.4 Result: ALLOW Hash: 2b6f5e3d9a11c7a2e9b1d4a7f8c2e8f3c7a2e9b1 Signature: MEUCIQb8....Ceoooe8goo...
USER LOGIN · Success
Authentication verified via cryptographic handshake
POLICY CHECK · Passed
All active rule sets evaluated deterministically
RISK EVALUATION · Low Risk
TRCP-Φκ coherence within bounds
ACCESS GRANTED · Resource: Financial.Report
Execution authorised. Cryptographic proof generated.
EVENT RECORDED · Aelthered Chronicle
Immutably sealed. ED25519 signed. Hash chained.
Who acted — actor identity cryptographically verified
What executed — full action context and payload
Which controls were active — governance state at execution time
What changed — before/after state with delta verification
Whether governance conditions remained structurally intact during execution

TauDIL Core Capabilities

TauDIL wraps any API-accessible LLM or existing core system. Deploys via a single HTTPS POST in hours. Intercepts prompts, evaluates against governance rules, and routes flagged verdicts via the UAE escalation engine — with zero external routing for air-gapped deployments.

C-01

TRCP-Φκ Coherence Scoring

Hallucination Prevention

Three lenses evaluate AI outputs: semantic entropy, contradiction detection, and goal overlap. TRCP-Φκ scores evaluate uncertainty, verify facts, and check alignment. Low scores route to Aelthered Chronicles with Kappa assignment before regulator audit export.

C-02

UAE Engine — Unified Assessment

Verdict Routing via Authority Matrix

Intercepts REVIEW or BLOCK decisions. Routes by domain, decision type, and authority level via matrix. Timers enforce review deadlines — triggering timed windows and auto-escalation to prevent bypass. Zero unreviewed outputs. SLA-bound accountability enforced.

C-03

Data Sovereignty

Air-Gapped Deployment

All prompts evaluate locally. No external cloud calls. Audit trails stay on-premises for air-gapped operations. Core Systems and Local AI Models connect to TauDIL via internal HTTPS — no external route exists by design.

C-04

Sealed Audit Logs

ED25519 + SHA-256 Chained

Every interaction feeds a gapless ledger: Rule Execution → Escalation Path → AI Output Hash → Compliance Assessment — all looped back cryptographically. Direct auditor access to query the chain for proof. Vendor bypass eliminated.

C-05

Universal LLM Wrapping

REST API · Hours to Deploy

Wraps any API-accessible LLM via REST. Intercepts prompts before they reach the model, governs outputs before they reach users. Integrates with Claude, Mistral, Cohere and enterprise platforms via HMAC-SHA256 webhooks.

C-06

Domain Liability Shield

Explainable AI Decisions

Converts raw LLM outputs to explicit verdicts with full evidence chains. Same input always produces the same output. EU AI Act Art. 13 transparency, GDPR Art. 22 explanation, Basel III / MiFID II audit trails — all structurally guaranteed.

TauDIL Governance Hub Claude Mistral Cohere SAP Workday Epic Guidewire Salesforce Air-Gapped Systems REST APIs HMAC-SHA256 Webhooks
Operation
Complexity
Typical Time
Notes
stage_batch (N entities)
O(N)
2–5s / 45 entities
Dominated by SYGON batch analyze
engine.analyze (batch)
O(N × lattice)
2–3s / 45 tokens
Pre-warms all tokens in one lattice pass
_sygon_signal (per entity)
O(1)
<5ms
Dict lookup only when _skip_analyze=True
AGL.decide (per entity)
<1ms
<1ms
Pure Python, no I/O
UAE Rule Evaluation
O(rules)
<500ms end-to-end
Deterministic. Same input = same output.
Constitutional Validation
O(1)
<1ms
Continuous runtime mode. Zero bypass risk.
τ

Deploy TauDIL in
your environment today.

One HTTPS POST. Hours to deploy. Deterministic governance from day one.

Request Architecture Demo Access Documentation Explore Rule Engine