CodeForge AI — Autonomous Multi‑Agent Coding Ecosystem (MVP Scaffold)

This repo bootstraps CodeForge AI: a multi-agent “AI engineering org” with a cinematic command-center UI, a realtime orchestration backend, and a deployable Docker setup (Hugging Face Docker Spaces compatible).

What’s included (MVP)

• Frontend (frontend/): React + Vite + Tailwind + Zustand + React Flow + Monaco Editor + Recharts + Framer Motion UI scaffold.
• Backend (backend/): FastAPI + WebSockets realtime event stream + in-memory “projects/agents/tasks” store (upgradeable to Redis/Postgres).
• Orchestrator (backend/codeforge/): agent registry, task routing, streaming event bus, basic consensus hooks.
• Docker: single container that serves FastAPI API + built frontend static assets (HF Docker Spaces friendly).

Local development

Backend (dev)

cd backend
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
uvicorn codeforge.main:app --reload --port 8000

Backend health: http://localhost:8000/api/health

WebSocket: ws://localhost:8000/ws

Frontend (dev)

cd frontend
npm install
npm run dev

Frontend dev server: http://localhost:5173

Set VITE_API_URL=http://localhost:8000 (optional; defaults to same-origin in production).

Production / Hugging Face Docker Space

Build and run locally:

docker build -t codeforge-ai .
docker run -p 7860:7860 -e PORT=7860 codeforge-ai

Then open http://localhost:7860.

HF Docker Spaces note: HF sets $PORT (usually 7860). This container listens on $PORT and serves both UI + API.

Next upgrades (planned)

• Redis-backed event bus + memory store
• Postgres projects/workspaces + PR simulation model
• Real agent model adapters (OpenAI/Claude/Gemini) with tool routing
• Auth (JWT) + RBAC for multi-project org
• CI pipelines + security scanning dashboards