About
Open-source governance infrastructure for AI agents. Developer preview.
Our Mission
AI agents are making real decisions in production. The audit layer underneath them doesn’t exist yet. We’re building it — open-source, composable, and designed to drop into the stack you already have.
Vorion gives developers the primitives to implement progressive trust — agents start sandboxed, advance tiers through recorded behavior, and every decision lands in a tamper-evident proof chain. Install from npm, wire it into your agent stack, ship agents you can audit.
Our Philosophy
Open Standards
The BASIS standard is open source because governance should be transparent. Anyone can inspect, critique, and improve the specification.
Progressive Trust
Trust is earned through consistent behavior over time. Our 8-tier model provides a clear path from sandbox to autonomy.
Human Oversight
Humans remain in the loop for high-risk decisions. AI augments human judgment; it doesn't replace it.
Observability
Every decision is logged. Every action is traceable. Full audit trails enable accountability and learning.
Open Source Commitment
Governance infrastructure should be open and inspectable. The core packages are Apache 2.0:
Apache 2.0
Apache 2.0
Apache 2.0
Apache 2.0