# Nova Autonomous LLM System

This directory scaffolds a next‑gen autonomous LLM stack suitable for container deployment using supervisord. It includes a gateway, a basic router, and a tools MCP server, and is designed to integrate with vLLM and external public endpoints.

Key components
- plane/api/gateway.py: OpenAI‑compatible proxy and route registry
- plane/orchestration/router.py: Task routing stub for multi‑backend selection
- plane/tools/mcp_server.py: Minimal MCP‑style tool registry and executor
- config/external_routes.yaml: External/public URL candidates (Cloudflare and local)
- requirements-nova.txt: Minimal runtime dependencies

Start locally
- vLLM: `bash elizabeth/deployment_configs/serve_vllm_no_steering.sh`
- Supervisord: `supervisord -c supervisord.conf` then `supervisorctl status`

Environment
- UPSTREAM_OPENAI_BASE: upstream OpenAI‑compatible base URL (default http://127.0.0.1:8000)
- EXTERNAL_ROUTES_FILE: path to routes YAML
- PORT: per‑service port overrides