Software learned to act. We're building the accountability.
Agentic Highway builds Vettd: the trust layer for the agentic supply chain. Software went from advising to operating in three years, and the discipline that kept software honest, the bill of materials, never made the jump to AI.
So we're building it: an open, vendor-neutral inventory of every agentic asset, graded on security and performance, with evidence that maps to the frameworks auditors already sign. Open source first. Zero egress always. Neutrality is the asset.
Operators from compliance, open source, and security.
The three disciplines this problem demands: regulated-industry product, open-source platforms at scale, and the security engineering to make evidence trustworthy.
Compliance & AI product — Fidelity, Pearson.
Open-source platforms — Open edX ecosystem.
Builds the scanner and evidence pipeline.
Advises the enterprise security roadmap.
Four commitments.
We grade agents; we don't sell them. The registry only works if no one can buy a grade.
MIT-licensed scanner, open evidence schema. Substrate, not another silo.
Scans run where your agents live. Nothing leaves the host until you decide it does.
NIST AI RMF, ISO 42001, EU AI Act, OWASP: we map to the frames auditors already sign.
Building agents? Auditing them? Talk to us.
We're partnering with security teams, auditors, and agent builders shaping the AI-BOM standard. Boston-based, remote-friendly, unreasonably fond of signed evidence.