A pioneering developer has unveiled a groundbreaking local AI companion project, leveraging cutting-edge technologies such as GWT, IIT proxy, ChromaDB hybrid retrieval, and Ollama fallback to create a sophisticated cognitive architecture.

This innovative system boasts 22 self-registering plugins that compete for attention through a Global Workspace Theory implementation, alongside an IIT consciousness proxy that provides a coherence signal. Additionally, the project incorporates modules for tracking live mood, energy, curiosity, attachment, and agency state, as well as homeostasis, self-modification, and intuition.

To ensure the system’s resilience, the developer has implemented per-module circuit breakers, a health monitor, and a 120-second watchdog. The Ollama fallback mechanism also kicks in if the primary Gemini system experiences downtime mid-session, minimizing disruption to the user.

The project’s personality model is based on the INFJ type, chosen for its seamless mapping to the architectural decisions already in place. The developer is now seeking feedback from the community on key aspects, including the GWT scoring implementation, IIT proxy framing, and hybrid retrieval weights.

The project’s code is available on GitHub, and the developer invites feedback and comments from the community to further refine this revolutionary AI companion.

Photo by Mikhail Nilov on Pexels
Photos provided by Pexels