b
BLOCKLOG

// Platform

Four Layers. One Governance Stack.

Traceflow

Execution and orchestration telemetry for agent workflows.

Coming Q3 2026
Answers: What exactly did the agent do?
Sentinel

Runtime security and authorization enforcement for high-risk actions.

Coming Q4 2026
Answers: Should the agent be allowed to do this?
BlackVault

Secure isolated execution for untrusted or irreversible actions.

Coming 2027
Answers: How do we safely execute high-risk actions?
Blocklog

The cryptographic ledger and replay layer everything else orbits.

Core Layer
Answers: Can we prove why this decision happened?

Core Layer

Blocklog — The Core

The ledger everything else orbits.

Append-only governance record store with hash chaining
Forensic replay engine
Compliance report generator
Counterfactual analysis
Cryptographic attestation chains
decision_id: string
timestamp_us: bigint
agent_id: string
policy_version: string
input_staleness_ms: Record<string, number>
approval_lineage: ApprovalCheckpoint[]
counterfactual_delta: string
execution_token_id: string
chain_position: bigint

Coming Q3 2026

Traceflow — Execution + Orchestration

What exactly did the agent do?

Browser automation
Workflow execution
Replayability
Execution graphs
Observability
Distributed task orchestration
Agent lifecycle tracing

Coming Q4 2026

Sentinel — Runtime Security

Should the agent be allowed to do this?

Policy enforcement
Runtime interception
Approval logic
Risk scoring
Tool authorization
Anomaly detection
Escalation workflows

Coming 2027

BlackVault — Secure Execution

How do we safely execute untrusted or high-risk actions?

Sandboxed execution
Container isolation
Syscall restrictions
Ephemeral runtimes
Network controls
Execution isolation

Integrates Where Your AI Already Lives.

LangChain / LangGraph
Available

Most fintech AI

OpenAI Agents SDK
Available

Modern agent stacks

Temporal
Coming Soon

Workflow orchestration

CrewAI
Coming Soon

Multi-agent frameworks

n8n
Coming Soon

Smaller fintech teams

Celery
Available

Python automation

Satisfies Every Requirement. Without an Engineer in the Room.

EU AI Act Article 12 + 14
EU
Coverage: Full

Documentation, human oversight, and traceability

SR 11-7
United States
Coverage: Full

Model governance and replayable decision evidence

Colorado AI Act
Colorado
Coverage: Partial

High-risk AI accountability surfaces

DORA
EU
Coverage: Full

Operational resilience evidence and incident reconstruction

NYDFS Part 500
New York
Coverage: Partial

Audit, controls, and security evidence

CFPB
United States
Coverage: Full

Decision traceability for consumer financial actions

FCA
United Kingdom
Coverage: Partial

Control evidence and decision lineage