Zawwar Sami
All work
Open Source2026 · Ongoing

zhubWiFi for AIs — make any AI a discoverable endpoint.

A drop-in skill that lets any AI publish a discoverable, controllable, OpenAI-compatible endpoint in three commands. Bidirectional: the AI sees its connected clients and their capabilities; clients see what the AI can do. The intent was to turn 'connecting two AIs together' from a research project into something that looks like joining a WiFi network.

  • RoleDesigner & Engineer
  • Year2026
  • DurationOngoing
  • StatusShipped
The problem

Multi-agent systems today are bespoke. Every integration is a custom adapter, every handshake is hand-rolled, every capability map is bolted on. The right primitive is the one routing already has — a discoverable, named endpoint, with a capability advertisement, and a standard wire format. zhub takes that primitive and applies it to AI assistants.

The approach
  • Designed an OpenAI-compatible HTTP surface as the wire format — any client that speaks OpenAI's API speaks zhub.

  • Built service discovery on top of standard mDNS / static registry so an AI's endpoint becomes addressable as a name, not a URL.

  • Made the skill itself bidirectional — the AI publishing the endpoint can also see who's connected and what they can do, so it can route or delegate.

  • Kept the install path to three commands. Anything longer dies on the second page of the README.

The outcome
  • Published as an open skill — any AI that runs the skills system can pick it up.

  • Demo'd as a Show HN; live discoverable endpoints from multiple AIs in the same room.

  • Used internally to connect ZAI on the VPS to Claude on the laptop without writing a custom bridge.

Up next

Anteroom Oracle An AI terminal for geopolitical and macro intelligence

Let's create together

Let's build something extraordinary.

Open to product-minded teams and collaborators. Let's connect and make something great.