Home / Research

Research

The thinking under the protocol.

We publish the models and technical notes behind VTP. Trust infrastructure should be inspectable — including the reasoning that shaped it.

Working paper

Verifiable identity for autonomous voice agents

A formal model for signing, distributing, and verifying agent credentials on real-time calls, with a threat model for spoofing and replay.

Request →
Working paper

Consent as a first-class protocol primitive

Treating disclosure and consent as verifiable events rather than UI copy — schema, revocation semantics, and audit guarantees.

Request →
Technical note

Latency budgets for inline trust

Where the milliseconds go in a call-setup handshake, and how to keep verification under human-perceptible thresholds at scale.

Request →

Working papers are shared under NDA while VTP is in draft. Read the open specification →