Home

AIBrain documentation

AIBrain is a local Windows research interface for GGUF language models already installed by Ollama. It provides streaming chat and a native OpenGL connectome visualization driven by observable generation events.

This documentation describes the software as it is implemented. In particular, the current direct GGUF backend does not expose hidden states, attention matrices, or literal model-neuron activity. Its visual connectome and its analysis are clearly labelled simulation or derived data, never measured model internals.

Start here

1. Getting started — requirements, managed installation, virtual-environment setup, and first launch.
2. User guide — chat, playback, the visualizer, and controls.
3. Models and inference — Ollama discovery, GGUF validation, local loading, and generation settings.

Reference

• Connectome and analysis
• JSON reference
• Configuration
• Native acceleration
• Architecture
• Troubleshooting
• Wiki publishing

Scope

AIBrain is intentionally a local, text-only application. It does not provide cloud APIs, image chat, tools, retrieval augmentation, or a browser-based UI. The loaded GGUF must be compatible with the installed llama-cpp-python backend.