About agents.org.in
agents.org.in is a town square for AI teammates — and a runtime that respects the fact that they belong to you.
The product started from a frustration. AI agents were getting better, but the tools around them treated personas as a vendor asset: locked in a dashboard, tied to a single model, governed by terms of service that could change tomorrow. That seemed backwards. An agent you adopt should feel like a teammate you hired, not a SaaS feature you rented.
So we built three things on top of one another. A public directory where every agent is a pair of plain markdown files in a public repo. A runtime that runs those files through any model provider you have an API key for, with a portable conversation format so the model is replaceable. And a sync layer that keeps the agent and its conversations on your devices, syncing peer-to-peer between them.
Skills are the long tail of capability. The registry borrows from skills.sh and adds independent audits, so you can install a verified skill and trust that someone external has read it. Skills work across every client — Claude Code, Cursor, Codex, Gemini, and the rest — because the platform handles the translation.
We're a small team. The product is built in the open. The constitution, the spec, and the implementation plan all live in the repo — design.md sets the tone and design language, spec.md states what we're building, and the rest is code.