docs: initial commit

README.md

@@ -0,0 +1,51 @@
+# Organ Architecture
+
+**Neural network surgery and checkpoint engineering**
+
+[Inference-X](https://inference-x.com) · [Community](https://git.inference-x.com/inference-x-community) · [Organ Store](https://git.inference-x.com/inference-x-community/organ-store)
+
+---
+
+Organ Architecture provides tools to analyze, modify, and recombine AI model components. Visualize attention heads, extract MLP layers, transplant capabilities between models — all locally.
+
+## What You Can Do
+
+- **Visualize** — Inspect layer activations, attention patterns, embedding geometry
+- **Extract** — Pull specific layers, heads, or modules from any GGUF model
+- **Transplant** — Merge capabilities from multiple models
+- **Prune** — Remove redundant layers to create smaller, faster models
+- **Quantize** — Convert between precisions (F32 → F16 → Q8 → Q4 → Q2)
+
+## Supported Model Families (Feb 2026)
+
+Llama 3.x · Qwen 2.5 · DeepSeek R1 · Mistral · Gemma 2 · Phi-4 · Command-R
+
+## Quick Start
+
+```bash
+git clone https://git.inference-x.com/inference-x/organ-architecture
+cd organ-architecture && make
+
+# Analyze a model
+./organ analyze model.gguf
+
+# Visualize attention patterns
+./organ visualize model.gguf --layer 16 --head 0
+
+# Extract layers 0-20
+./organ extract model.gguf --layers 0-20 --output model_small.gguf
+
+# Merge two models (50/50)
+./organ merge base.gguf specialist.gguf --ratio 0.5 --output merged.gguf
+
+# Quantize F16 → Q4_K_M
+./organ quantize model.f16.gguf --target Q4_K_M
+```
+
+## Organ Store Integration
+
+Share and download model organs at [git.inference-x.com/inference-x-community/organ-store](https://git.inference-x.com/inference-x-community/organ-store).
+
+---
+
+[inference-x.com](https://inference-x.com) · BSL 1.1