smart-dispatch

Token optimization is achieved by mapping task requirements to Opus, Sonnet, or Haiku routing. The system evaluates task complexity and recommends the most cost-effective model, providing a comparative token-economy estimate versus defaulting to Opus.

Model routing assigns tasks to appropriate language models based on complexity classification. It analyzes whether work is complex, intermediate, or simple, then selects Opus, Sonnet, or Haiku respectively, producing a clear model selection rationale for each request.

Parallel dispatch planning is supported when parts of a task can run independently. The system identifies opportunities to split work across multiple models, recommending parallel execution to optimize processing for drafting, reviewing, analyzing, or transforming documents.

Smart-dispatch applies to legal, software, and general productivity workflows. It recommends cost-effective models by mapping task requirements, covering drafting, reviewing, analyzing, and transforming documents across these specific professional domains.

Model routing may be unnecessary when tasks strictly require the highest reasoning capabilities of a single large model regardless of cost. If work cannot be split for parallel execution and demands consistent Opus-level processing, routing provides minimal token savings.