Skip to content

Narwhal documentation

The Narwhal logo, a black narwhal with a teal spiral tusk above the wordmark

Narwhal is a Python framework for disaggregated LLM inference. It assigns prefill and decode roles across dual-capability engines as demand changes. Read the README for an overview and the product contract.

Run Narwhal

Reference

  • Core concepts


    Request flow, topology, role control, failure, and state contracts.

  • Configuration


    Fleet fields, deployment inputs, and validation rules.

  • CLI


    Commands, options, results, and exit codes.

  • HTTP API


    Completion, inspection, live-state, and lifecycle routes.

  • Telemetry and artifacts


    Journal, profile, metric, failure, and compatibility contracts.

Read the contributor guide, code of conduct, and security policy before contributing.