Installation¶
Prerequisites¶
- Python 3.13+
- uv for Python package management
- Graphviz for schema diagram rendering
- Access to a Hawk deployment (API server URL + OAuth2 credentials)
Install the CLI¶
From GitHub:
bash
uv pip install "hawk[cli] @ git+https://github.com/METR/hawk#subdirectory=hawk"
From source:
bash
git clone https://github.com/METR/hawk.git
cd hawk/hawk
uv pip install -e ".[cli]"
Configuration¶
Set these environment variables before using the CLI, or put them in a .env file:
bash
export HAWK_API_URL=https://hawk.example.com
export INSPECT_LOG_ROOT_DIR=s3://my-bucket/evals
| Variable | Required | Description |
|---|---|---|
HAWK_API_URL |
Yes | URL of your Hawk API server |
INSPECT_LOG_ROOT_DIR |
Yes | S3 path for eval logs |
HAWK_LOG_VIEWER_URL |
No | URL for the web log viewer |
HAWK_DATADOG_EVAL_SET_DASHBOARD_URL |
No | Datadog dashboard URL for eval sets |
HAWK_DATADOG_SCAN_DASHBOARD_URL |
No | Datadog dashboard URL for scans |
HAWK_MODEL_ACCESS_TOKEN_ISSUER |
No | OIDC issuer URL for authentication |
HAWK_MODEL_ACCESS_TOKEN_CLIENT_ID |
No | OIDC client ID |
HAWK_MODEL_ACCESS_TOKEN_AUDIENCE |
No | OIDC audience |
Authentication¶
bash
hawk login
This starts an OAuth2 Device Authorization flow. Follow the on-screen instructions to authenticate.
Run your first eval¶
bash
hawk eval-set examples/simple.eval-set.yaml
hawk logs -f # watch it run
hawk web # open results in browser