| # Startup script for Tiny Narrator on Hugging Face Spaces free CPU. | |
| # Model inference endpoints are configured through environment variables. | |
| set -eu | |
| echo "[tiny-narrator] Starting Tiny Narrator app on ${GRADIO_SERVER_NAME:-0.0.0.0}:${PORT:-${GRADIO_SERVER_PORT:-7860}}" | |
| if [ -n "${LLAMA_CPP_BASE_URL:-}" ]; then | |
| echo "[tiny-narrator] Reader brain endpoint configured at ${LLAMA_CPP_BASE_URL}" | |
| else | |
| echo "[tiny-narrator] LLAMA_CPP_BASE_URL is not configured; reader brain will use deterministic fallback." | |
| fi | |
| exec python app.py | |