LONG ARC STUDIOS
Building the accountability layer for autonomous systems.
THE THESIS
AI agents will manage more value than humans within a decade. The question is not whether they will act autonomously. It is whether we can prove what they did when something goes wrong.
Permission is not proof. Logs are self-reported. The gap between "the agent did something" and "we can defend what the agent did" is where liability lives.
mdash closes that gap.
THE ECOSYSTEM
mdash
Protocol
The cryptographic attestation protocol. Warrants, attestations, physics validation, checkpoint surety. Accountability infrastructure for the agentic economy.
V3.1.0
Kiwi
Runtime
The accountable runtime for autonomous AI agents. Bounded execution via seccomp-bpf, Agent Profiles for human context, and insurance-grade audit trails.
Q2 2026
TRACTION
V3.1.0
shipped
612
tests passing
20
attack vectors documented
Open
Design Partner Program
Active
private beta
CONTACT