economy-mode

Reduce LLM operating costs by switching Layer 3 task-aware auto-selection to cheaper alternatives using economy mode. It applies an economy model table that preserves your Layer 0, 1, and 2 explicit intent while lowering compute expenses for coding, documentation, planning, triage, and review workloads.

Switch language model routing to cheaper alternatives for a single session by using trigger phrases like “save costs” with your agent. This activates economy mode immediately, applying lower-cost model mappings to your current tasks while keeping your explicit preferences intact.

Yes, you can persist cost optimization settings across sessions by writing to the `.squad/config.json` file. Persistent activation ensures economy mode applies automatically to future agent spawns, continuously routing Layer 3 tasks to lower-cost alternatives.

Economy mode handles agent spawning by applying a defined economy model table for Layer 3 model selection. When active, spawned agents include a 💰 indicator in their messages, confirming that lower-cost alternatives are being used for task outputs.

No, applying cost optimization settings never overrides your Layer 0, 1, or 2 explicit model preferences. Economy mode strictly targets Layer 3 task-aware auto-selection, ensuring your highest-priority configuration and explicit intent remain fully intact.

LLM cost optimization mode supports common workloads including coding, documentation, planning, triage, and review tasks. It shifts the Layer 3 auto-selection for these specific outputs to cheaper alternatives, reducing overall compute costs without losing task intent.