🚀 Decodifier v0.2.0 — The Pattern Compiler Release
This release is the first major step toward code compilation at scale for AI-generated software.
LLMs don’t need to write code. They write intent.
Decodifier compiles that intent into working software — without the model ever reading the codebase.
✨ What’s New
🧱 Pattern Engine (≈98 patterns)
📦 Spec Loader + Validator
⚙️ FastAPI Backend Generation
🔁 Router Autowiring
🧪 LLM-Safe File Read/Write
📊 Dashboard + Build History
🧠 Architecture → Patterns → Specs → Code workflow
🎯 Why this matters
Prevents architecture drift
Eliminates hallucinated imports
Reduces LLM context load by 10–50x
Moves AI coding from “wizard” to compiler stage
A new layer in the AI stack
🛠️ Getting Started
pip install -r requirements.txt
uvicorn engine.app.main:app --reload
open http://localhost:8000/dashboard
Add a spec → Click Generate from Specs → watch the backend evolve.
🗺️ Roadmap
v0.3.0: CLI, CRUD emitters, schema adapters
v0.4.0: LLM tool adapters (OpenAI, Gemini, Claude)
v1.0.0: Full end-to-end compiler chain
📄 License
AGPL-3.0 — to ensure improvements remain open when used as hosted services.
Commercial licenses available for closed-source / enterprise deployment.