a federated, pluggable LLM memory system
A single llm-wiki is good. But you will eventually want more than one. And then you'll want them to talk to each other. That's what this is.
federated-llm-wiki should let you deploy vaults in minutes, connect them in a clean DAG, plug in agents and UI components, and scale from a single personal wiki to a network of federated knowledge bases — covering completely unrelated topics — under one roof.
Not centralized — So, no global authority. Not decentralized — So, no trustless chaos. Federated: each vault governs its own knowledge; connections are explicit, directional, and acyclic.