BoggersTheAI — TS-OS (Thinking System Operating System)

A local-first, living cognitive runtime.

Instead of a static LLM, BoggersTheAI runs a continuous wave-propagation graph where concepts are nodes with activation/stability, and thoughts emerge from constraint resolution and wave dynamics.

The graph is alive in the background — propagating, relaxing, pruning, resolving tensions, and spawning emergent structure — while remaining fully observable and controllable.

Local. Self-improving. Multimodal. No cloud required.

---

Quick Start

git clone https://github.com/BoggersTheFish/BoggersTheAI.git
cd BoggersTheAI

pip install -e ".[all]"

# Start Ollama + recommended models
ollama pull llama3.2
ollama pull nomic-embed-text

# Launch the CLI
boggers

Type help in the prompt for commands.

Launch the web dashboard:

dashboard-start

---

Core Philosophy

Thought is not token prediction. Thought is the stable configuration of constraints under wave propagation.

The system maintains a living graph that evolves continuously through cycles of:

Propagate → Relax → Prune → Merge → Detect Tension → Evolve

LLMs are used only as a synthesis tool when needed. The real intelligence lives in the graph dynamics.

---

Key Features

• Living Wave Graph with continuous background evolution
• Self-Improvement via trace → QLoRA fine-tuning + hot-swap
• Multimodal (voice in/out, image understanding)
• Rich observability — CLI, TUI, and beautiful FastAPI dashboard with live graph viz
• Tool use + external adapters (web, RSS, Wikipedia, etc.)
• Thread-safe with proper mode locking between user queries and autonomous cycles

---

Part of the TS Ecosystem

• BoggersTheCIG — Self-evolving epistemic knowledge graph engine
• TS-Core — Lightweight graph dynamics kernel (Python + Rust acceleration)
• GOAT-TS — Theoretical foundation
• TensionLM / bozo — Experimental models using tension graphs

---

Next Steps

• Run your first queries in the CLI
• Explore the live graph in the dashboard (/graph/viz)
• Watch the wave cycle in real time
• Trigger self-improvement with improve command

Full technical deep-dive and configuration reference is preserved in the previous long README version while we iterate on clarity.

---

Status: Actively developing (v0.5+ wave)

License: MIT