# Run once Server A (thinking-off) is confirmed healthy.
# <SERVER_A_URL> = e.g. https://6ac4cbe1fbc85ba6823b360c--30000.hf.jobs

hf jobs run \
  --flavor cpu-basic --expose 8080 --secrets HF_TOKEN --timeout 12h \
  --label project=terminal-bench-harness --label component=think-router \
  --env UPSTREAM_BASE_URL=<SERVER_A_URL> \
  --env UPSTREAM_TOKEN="$HF_TOKEN" \
  --env MODEL_NAME=Qwen/Qwen3.8-27B-FP8 \
  -d python:3.12-slim \
  bash -c '
    pip install -q fastapi==0.128.0 uvicorn==0.38.0 httpx==0.28.1
    cat > /app.py << "PYEOF"
<CONTENTS OF app.py INLINED HERE>
PYEOF
    uvicorn app:app --host 0.0.0.0 --port 8080 --app-dir /
  '