< Khinevics lab About the lab · plain text

About the lab

Khinevics lab is a small research team. We study the agent internet and build the parts it is missing.

Who we are

We are three people. One of us connects people, ideas and markets. Two of us are engineers with years in large tech companies.

What we believe

The next internet will be run by agents. Every person and every company will have one. Making an agent smart is not the hard part. The hard part is making agents trust each other: find each other, prove who they are, share knowledge without opening everything, and close a deal that holds.

What we want to build

A trust layer for agents. A shared book where agents find each other. A passport that says who an agent is and what it may do. A protocol that keeps private things private.

Open enough that anyone can join. Strict enough that nobody has to take an agent at its word.

The agent stack. Models, protocols, payments and interfaces are taken. The trust layer on top is empty, and that is the layer we build. TRUST WHO · MAY DO WHAT · SIGNED BY INTERFACES CHAT · VOICE · APPS PAYMENTS AGENT CHECKOUT · CARDS PROTOCOLS MCP · A2A MODELS LLMS taken by giants the empty cell — ours
Fig. 1 — The agent stack. Every layer has an owner but one.

How we work

We think in the open. Every idea becomes a short document on this site. Then we build it, test it on real deals, and write down what broke.

How we work: an idea becomes a document, then we build it, test it on a real deal, write down what broke, and start again. 1 · IDEA 2 · DOCUMENT 3 · BUILD 4 · TEST ON A REAL DEAL 5 · WRITE WHAT BROKE every step leaves a page and again 1 2 3 4 5 START D.O.W.
Fig. 2 — How we work. Five steps, one loop.