Loading AGEI...
AGEI is AI Governance Evidence Infrastructure
AI systems submit evidence to server-side governance gates. AGEI evaluates policy, produces outcomes, records receipts, links evidence, and supports later audit verification.
AI systems do not approve themselves. They submit evidence. AGEI evaluates policy server-side and records the outcome as verifiable evidence.
Start here if you are new to AGEI. These guides explain the core idea: policies define governance, gates enforce it, outcomes become receipts, and receipts form verifiable evidence chains.
What AGEI is and why it exists
The enforcement-backed evidence infrastructure. Learn about the core guarantee: no self-certification, server-side evaluation, and four-outcome gate model.
Get AGEI running quickly
Set up Supabase, run migrations, configure environment, and evaluate your first gate. Complete setup in 5 minutes.
The AGEI evidence model: Policy → Gate → Evaluation → Receipt → Evidence → Vault → Audit Pack → Verification
The 10 default policies and gates that make AGEI evidence canonicalized, hashed, signed, and verifiable.
Policies define governance intent. Gates are operational enforcement points.
Coreapprove • deny • escalate • inspect - The four governance outcomes.
CoreAtomic evidence records with cryptographic integrity. Core of AGEI evidence spine.
CoreHow receipts link into a chain of custody. Trains on, validates with, deploys, infers with.
Lifecycle PackSealed evidence for retention. Immutable capsules with reconstruction proofs.
CoreLazy Capsule Materialization. Materialize evidence only when auditors need it.
CoreDownload the Python SDK and get support for integration
Get the official AGEI Python client library
Get help with SDK integration
Practical guides for integrating AGEI into your AI systems
RESTful API reference and examples
Python client for ML workflows
AGEI Core is required. Extension packs add specific capabilities depending on your use case.
Foundation for cryptographic AI governance evidence.
How AGEI ensures cryptographic integrity of evidence
Canonicalize → hash → compare. Production-grade recursive key sorting matches JavaScript byte-for-byte.
Technical architecture for buyers, auditors, and contributors
Deploy, test, and monitor AGEI in production