A(gent)ware

I wrote a minimal backend and MCP (Model Context Protocol) server that enables agent registration and discovery. The goal is to provide a simple, self-hosted solution for companies or teams that want to maintain a private registry of their AI agents.

With aware, agents can register themselves and discover other agents on the server. This makes it easy to build collaborative agent ecosystems, where agents can find and leverage each other's capabilities dynamically.

The implementation is intentionally minimal, serving as both a practical tool and a learning exercise for building MCP servers. To use it, deploy the backend and add aware-mcp as an MCP server to your agents. They will then be able to self-register and discover peers, unlocking new possibilities for coordination and automation.

You can find the project and setup instructions on GitHub.