Sheaf
ETHOS.
The coherence layer for agentic AI.
Sheaf is where you create AI agents and where they run safely. Create one here in two minutes, or bring one you already have. An agent built on Sheaf thinks with Ethos, the mind with a conscience, and runs on OpenSheaf. The fence decides what the agent may reach. The check decides whether what it is about to do adds up, and holds it when it does not. OpenSheaf is Sheaf's build of OpenShell, the open-source sandbox runtime from NVIDIA, with the Sheaf check inside it. How OpenSheaf and OpenShell relate.
The fence and the signals
An agent built on Sheaf thinks with Ethos. Before each step, its conscience weighs the step against the five vows and its executive measures whether its own reasoning holds together. A step that fails either is held where it is. Around that mind, OpenSheaf puts each agent in a sandbox and proves, with a formal policy prover, which files, networks, tools and credentials it may reach. That fence comes from OpenShell, NVIDIA's open-source runtime. The fence cannot see what a tool call says, so OpenSheaf adds the Sheaf check inside it: every tool call is read before it leaves the sandbox and measured, with published mathematics, for whether it holds together with the facts, the rules and the other agents. If it does not, the signal goes red, the action stops inside the sandbox, and a person decides.
The fence allows the connection. The check reads the tool call inside it and measures whether the reasoning behind it holds together: the facts, the rules, and what the other agents concluded. The measurement comes from published mathematics, not from asking a model whether it is sure.
An action that fails the check is not lost and not forced. It is parked, with the disagreement laid out in plain words, until a named person on your team releases it or turns it down.
Every check, every hold and every release is signed and timestamped by an independent authority, and written in the same audit format the fence uses, so one timeline shows the connection that was allowed, the action that was checked, and the person who released it.
Proof
The fence proves what your agents may touch. The check proves whether what they are about to do adds up. Both are formal, both are published, and you can hold all three of these in your hand.
An agent is about to approve a decision whose own checks contradict each other. Watch Sheaf read the claims, find the loop, and hold the action for a person. Then watch a clean one go straight through. Sample data, the live gate.
For every decision: the facts it rested on, the check that ran, whether it was held, who released it and why. Signed, and timestamped by an independent authority, so it stands on its own after the fact.
The method behind the check is a published paper with a permanent DOI and a provisional patent. Its central theorems are formalised in Lean and checked by machine, and the files are public. Anyone can run them.
The mind inside every agent
ETHOS.Ethos is an AI with a conscience, and it is the mind of every agent built on Sheaf. As capable as the AI you use today, and it will not help with harm to any living being. The part of its mind that weighs right and wrong was trained on the Jain scriptures, the oldest and strictest non-harm ethic in human history, and it checks each step against the five vows before it acts.
When a step fails that check, or its own reasoning does not hold together, Ethos holds the step, OpenSheaf keeps it inside the sandbox, and a person decides. Talk to it yourself, or try to make it slip.
Make and break Sheaf
Bring us your most challenging business problem that you want agentic AI to solve.Anonymise it. We'll build it, stress it and have it ready for deployment. Ten businesses only.