README.md

---
title: Codette · Orchestrated Reasoning
emoji: 🕸️
colorFrom: purple
colorTo: indigo
sdk: gradio
sdk_version: 6.20.0
app_file: app.py
short_description: Codette's real orchestrated reasoning pipeline
python_version: "3.12"
startup_duration_timeout: 1h
pinned: false
models:
  - Raiff1982/Codette-Reasoning
  - NousResearch/Meta-Llama-3.1-8B-Instruct
tags:
  - codette
  - multi-perspective-reasoning
  - orchestrator
  - lora
  - llama-3.1
---

# 🕸️ Codette — the orchestrated system

This Space runs **Codette's real reasoning pipeline**, not raw adapters. The
faithful llama.cpp orchestrator's model layer is swapped for a transformers+peft
backend so it can run on ZeroGPU; everything above the model primitive —
routing, the permanent behavioral locks, the intellectual-integrity layer
(complexity + role matching), constraint enforcement, and self-correction — is
the actual Codette code, unchanged.

- **Auto mode** routes each query to the perspective Codette judges best.
- Or force one of the eight perspectives (Newton, DaVinci, Empathy, Philosophy,
  Quantum, Consciousness, Multi-Perspective, Systems Architecture).

**Faithful, with honest caveats for this hosted demo:**
- Codette's **foundational cocoon memory** (identity, honesty, the people and values
  she carries) is loaded read-only from her `CodetteBrain` bucket and folded into
  every response. New cocoons aren't written back (read-only, stateless per request),
  so there's no cross-session learning here.
- Generation runs on transformers+peft (Llama-3.1-8B) instead of the local
  OpenVINO/llama.cpp backend; the reasoning logic is identical.

Built solo by **Jonathan Harrison** (Raiff1982). Codette is accepted for
publication in *Scientific Reports* (Nature Portfolio).
Model: <https://huggingface.co/Raiff1982/Codette-Reasoning>