What if your personal AI assistant ran on your own hardware, stored its memory in files you control, and remained accessible to the tools you already use?

In this session, I’ll demonstrate the personal agent I built and take you through its architecture. It runs at home using local language models and stores its long-term knowledge as portable, human-readable text rather than hiding it inside a proprietary platform.

We’ll see how other applications can securely access the agent through an open protocol, allowing different tools and AI clients to work with the same memory and capabilities without exposing the home network or surrendering control of the underlying data.

Along the way, we’ll explore the decisions behind local inference, persistent memory, tool integration, identity, and secure remote access.

You’ll leave with a practical blueprint for building a personal agent that is local, connected, portable, and—most importantly—under your control.