The introduction video of EasyNet.Run

🚀Let AI use tools and context wherever they are — just like you use the internet.

Your AI agent can write code, plan a trip, and debug a server.

It still can't use the model your agents trained on your teammate's GPU. It can't bring yesterday's important AI conversation into your teammate's app. Not easily, not safely, not reliably.

Not because the AI isn't smart enough, but because there's no standard way to reach them: no address, no permission, no record of what happened.

The internet solved this for people decades ago. You use a website without owning its server. Agents don't have that yet.

So for the past 7 months, alongside my PhD at NUS, I've been building it: EasyNet (easynet.run).

→ Every tool and piece of context gets an address, wherever it lives → The owner decides who can use it and what comes back; models, code and data never leave their machine → Every call comes back with a signed receipt

MCP is how an agent plugs into a tool. EasyNet is how it reaches tools and context on someone else's machine. They work together.

A protocol, a device runtime, and SDKs in six languages. Open source, now in technical alpha.

If you also want agents to use tools and context from anywhere, the way people use the internet, whether as an investor or as someone who wants to build it with me, let's talk. 👇