model-router

Route tasks to the right AI model by classifying every incoming message to determine whether Opus for planning and debugging or Sonnet for execution and formatting is required. This ensures the optimal level of thinking is applied automatically.

Automatic task orchestration for AI agents is the process of routing tasks to subagents based on complexity classification. It enforces strict tier boundaries to delegate model-appropriate work like architecture planning to Opus and simple edits to Sonnet.

Automatically assign Opus for planning and Sonnet for execution by providing a task to a model router. It performs on-every-message classification to delegate work to the appropriate tier, ensuring strategy tasks use Opus and repetitive editing uses Sonnet.

Yes, you can use model routing for debugging sessions and architecture planning. The router classifies these complex tasks and directs them to the Opus tier, while execution and formatting tasks are routed to Sonnet for efficient processing.

The best way to automate subagent delegation for parallel AI workflows is using a model router that implements strict tier boundaries. It classifies complexity on every message and orchestrates subagents to handle execution-tier processing concurrently.

You should not use complexity classification for model routing when tasks do not fit clearly into planning or execution tiers, or when strict tier boundaries would prevent necessary fluid interaction between model tiers during complex debugging sessions.