Skip to main content
Attestix
v0.2.4

Now Open Source

Attestix

The EU AI Act takes effect August 2, 2026. Non-compliant organizations face fines up to EUR 35 million or 7% of global revenue. Attestix is like TurboTax for AI compliance: it automates the documentation, identity verification, and audit trails your AI agents need to stay legal. Install once, generate cryptographic proof of compliance on every run.

Star on GitHub

Open source. Apache 2.0 license.

9 Modules47 ToolsEU AI Act ReadyOpen Source

Architecture

9 Modules, One Coherent Stack

Identity
DIDs
Agent Cards
Who is this agent?

Identity Layer

did:key and did:web resolution with UAIT bridge connecting MCP, A2A, and DID ecosystems.

Credentials
Delegation
Compliance
What can this agent do?

Trust Layer

W3C Verifiable Credentials with Ed25519 signatures, UCAN delegation chains with scoped permissions, and regulatory compliance tooling for EU AI Act workflows.

Provenance
Reputation
Blockchain
What did this agent do?

Observability Layer

Hash-chained provenance tracking, reputation scoring from verified actions, and blockchain anchoring to Base L2 via EAS for tamper-proof audit trails.

Technology Stack

Standards, Clients, and Infrastructure

Standards & Protocols
W3C VCs
W3C DIDs
UCAN
Ed25519
EU AI Act
MCP
A2A
IETF
JSON-LD
JWT
DID:key
EAS
MCP Clients
Claude Desktop
Cursor
Continue
Windsurf
VS Code
Blockchain
Base L2
Ethereum
EAS

Code Examples

example.py
# Create a verifiable identity for your AI agent
# Generates a UAIT (Unique Agent Identity Token) and DID

result = await create_agent_identity(
    display_name="compliance-auditor",
    source_protocol="manual",
    capabilities="audit, verify, report",
    description="Automated compliance auditing agent",
    issuer_name="Acme Corp",
    expiry_days=365
)

# Returns:
# - agent_id: "attestix:a1b2c3..."
# - did: "did:key:z6Mk..."
# - public_key: Ed25519 public key

Use Cases

Agent identity created
Compliance profile assessed
Credential issued and anchored

Compliance Pipeline

Identity creation to compliance assessment to credential issuance in one flow.

[2026-02-27 09:15:22] IDENTITY

Agent identity created. UAIT assigned.

[2026-02-27 09:15:24] PROVENANCE

Training data recorded. Hash chain extended.

[2026-02-27 09:15:27] CREDENTIAL

Verifiable credential issued. Ed25519 signed.

[2026-02-27 09:15:30] COMPLIANCE

Conformity declaration generated.

[2026-02-27 09:15:33] ANCHOR

Anchored to Base L2 via EAS. Tx confirmed.

Audit Trail

Hash-chained provenance tracking from identity to blockchain anchoring.

By the Numbers

0

MCP Tools across 9 modules

0

Tests with conformance benchmarks

0

Conformance test suites (W3C, IETF, UCAN, MCP)

Voices on AI Agent Trust

What People Are Saying

Researchers, founders, and industry leaders validating the need for verifiable trust infrastructure in AI agent systems.

Ready to add verifiable trust to your AI agents?