# ── LLM backend (one of these is required for inference.py) ────────────────── # Groq — free tier, fast. Get a key at console.groq.com GROQ_API_KEY= # HF Inference Router — requires HF account. Get a token at huggingface.co/settings/tokens HF_TOKEN= # ── Optional overrides ──────────────────────────────────────────────────────── # Override the auto-detected API base URL # API_BASE_URL=https://router.huggingface.co/v1 # Override the default model for the chosen backend # MODEL_NAME=Qwen/Qwen2.5-72B-Instruct # Override the Narada environment URL (default: live HF Space) # ENV_URL=http://localhost:7860 # ── Server config (only needed if running the server locally) ───────────────── # PORT=7860 # HOST=0.0.0.0 # WORKERS=1