Immutable on-chain audit trails for autonomous AI agents.
EU AI Act compliant across 6 blockchain networks.
Know Your Agent — before it acts on your behalf.
Off-chain logs can be modified. Databases can be wiped. On-chain logs are permanent by design — across 6 networks simultaneously.
Every action written to the blockchain. No admin key. No delete function. No exceptions. 6 networks, one SDK.
ZKAuditProof.sol — prove compliance without revealing sensitive data. 169 tests passing. Regulator-ready.
Satisfies Articles 9, 11, 12, 13, 14, 19, 26, 72, 73. Legal co-founder with EU AI Act expertise.
Full implementation of Trustless Agents standard. Only EU AI Act-aware ERC-8004 reference implementation.
EUAIActReporter.sol generates audit-ready compliance reports automatically. 127 tests passing.
4 on-chain tools via MCP protocol. TypeScript SDK with ElizaOS + LangChain integrations.
Every EU AI Act requirement mapped to a specific on-chain contract. Not a checkbox — an immutable evidence layer.
| Article | Requirement | AgentAudit Contract |
|---|---|---|
| Art. 9LIVE | Risk management system | AuditVault + Risk Scoring HIGH/MEDIUM/LOW |
| Art. 11LIVE | Technical documentation (AIBOM) | AIBOMRegistry.sol — CycloneDX 1.6 schema |
| Art. 12LIVE | Record-keeping for high-risk AI systems | AuditVault — immutable on-chain log per action |
| Art. 13LIVE | Transparency and information to users | Every action timestamped, publicly readable on-chain |
| Art. 14LIVE | Human oversight | Dashboard + on-chain override mechanisms |
| Art. 19LIVE | Logging requirements | Automatic per-action logging via AuditVault |
| Art. 26LIVE | Obligations for deployers | AgentRegistration + KYA fields enforced on-chain |
| Art. 50LIVE | AI content transparency | Decision Provenance + on-chain logs |
| Art. 72LIVE | Post-market monitoring | PostMarketMonitor.sol — continuous activity log |
| Art. 73LIVE | Incident reporting (15/10/2 day timelines) | IncidentRegistry.sol — automated reporting |
Drop-in integrations for the most popular AI agent frameworks. One SDK, any chain, any runtime.
Native @elizaos/plugin-agentaudit — automatic EU AI Act logging for every ElizaOS agent action. Zero config.
Callback handler for LangChain agents. Every LLM call, tool use, and chain step logged on-chain automatically.
4 on-chain tools via Model Context Protocol. Works with Claude, Cursor, and any MCP-compatible runtime.
@agentauditai/sdk — universal SDK for any TypeScript agent. Supports all 5 networks. Two lines to start logging.
Open-source core. Drop into any AI agent runtime in minutes.
Get full AgentAudit AI infrastructure — all 5 networks, all contracts, SDK + MCP server — completely free as a Founding Member pilot. Aug 2 deadline is not a suggestion.