Accountability Infrastructure
AGLedger Implementation
Deploy accountability infrastructure for your AI agents. Mandate-Receipt-Verdict lifecycle, Ed25519-signed audit trails, settlement signals, and EU AI Act compliance mapping — self-hosted on your infrastructure.
What is AGLedger?
AGLedger is the accountability layer for automated operations. Any process that speaks HTTP or MCP can participate — AI agents, RPA bots, CI pipelines, microservices.
The chain: Principal assigns a mandate. Performer delivers a receipt. Principal renders a verdict. AGLedger records everything — commitment, delivery, verdict, timeline, audit trail — in a hash-chained, Ed25519-signed vault.
Built by Tributary's founder. Open documentation at agledger.ai.
Deployment Modes
Standalone
Single-instance deployment for organizations getting started with agent accountability. One AGLedger database backing your internal agent fleet. Ideal for teams running 1-50 agents.
Gateway
AGLedger sits between your agent orchestrator and downstream services, intercepting mandates and receipts at the network boundary. Zero changes to existing agent code — accountability via configuration.
Hub (Federation)
Multi-org accountability across supply chains, partnerships, or regulated ecosystems. Each party runs their own AGLedger instance; the hub federates mandate chains and settlement signals across trust boundaries.
All modes deploy via Docker or Helm on your infrastructure. Air-gap capable — no external network dependencies at runtime.
Integration
REST API
Full lifecycle management via HTTP. Create mandates, submit receipts, render verdicts, query audit trails. OpenAPI 3.1 spec included.
TypeScript & Python SDKs
First-class SDKs for the two dominant languages in AI/ML. Type-safe mandate creation, receipt submission, and audit queries with built-in retry and error handling.
MCP Server
Model Context Protocol support lets LLM-based agents interact with AGLedger natively. Agents can create mandates, submit receipts, and query their own accountability records through MCP tools.
What AGLedger Records
Mandate Lifecycle
Full state machine: CREATED, ASSIGNED, IN_PROGRESS, DELIVERED, ACCEPTED, REJECTED, SETTLED. Every transition timestamped and signed.
Audit Vault
Hash-chained, Ed25519-signed entries. Tamper-evident by cryptographic construction, not organizational policy.
Settlement Signals
SETTLE or HOLD signals for payment platforms. Accountability-driven authorization for financial operations triggered by agent work.
EU AI Act Mapping
Compliance mapping across 11 articles. Dual Trail crosscheck (Declare vs Detect) for Articles 13 and 14 transparency and human oversight requirements.
Use Cases
Enterprise Agent Accountability
Every agent action gets a mandate, every result gets a receipt, every outcome gets a verdict. Your audit trail is hash-chained and Ed25519-signed — tamper-evident by design, not by policy.
Regulatory Compliance
AGLedger maps to 11 articles of the EU AI Act. Dual Trail compliance crosscheck (Declare vs Detect) gives regulators the evidence trail they need without custom reporting.
Supply Chain Orchestration
Federated mandate chains across organizations. When your agent delegates to a partner's agent, both sides maintain independent, verifiable accountability records.
Financial Operations
Settlement signals (SETTLE/HOLD) integrate with payment platforms. Agents do the work, AGLedger provides the accountability signal that authorizes — or blocks — financial settlement.
Engagement Model
Implementation scope: Scoped after a discovery call — deployment mode, integration points, and team enablement
Timeline: 4-8 weeks typical (Standalone), 6-12 weeks (Gateway/Hub)
AGLedger license: $8K perpetual per database instance + $4K/yr optional support
Typical buyer: CTO, VP Engineering, Head of AI/ML, Chief Architect
No per-agent fees, no per-transaction costs, no usage metering. The implementation engagement covers deployment, integration, configuration, and team enablement.
Frequently Asked Questions
Ready to Deploy Accountability Infrastructure?
Book a discovery call to discuss your agent architecture, deployment requirements, and compliance needs.