ecomode

To reduce Claude Code operational costs, use model routing skills like ecomode that automatically prefer cheaper model tiers like Haiku or Sonnet and escalate only when necessary, preserving essential outcomes while lowering token usage.

Token-efficient model routing in Claude Code is the process of automatically directing requests to lower-cost model tiers when safe to do so. It escalates to higher tiers only when necessary to complete the task without compromising results.

You can route Claude Code tasks to cheaper models by prefixing your task commands with eco, such as executing eco ralph or eco autopilot. This applies cost optimization rules across delegation decisions automatically.

Yes, ecomode works as a modifier with existing execution modes in Claude Code. You can combine it with modes like eco autopilot or eco ralph to preserve task autonomy while reducing cost and token usage.

You should escalate a task to a higher Claude model tier only when the cheaper models cannot safely handle the request or when necessary to preserve essential outcomes. Ecomode applies this escalation logic automatically during delegation.