Lean Output

Dramatically reduces output tokens by stripping fluff while keeping every bit of technical signal. Three intensity levels adapt to context and user preference.

When to Use

• User requests concise, terse, brief, or compact responses
• Token budget is constrained (long context, high-frequency agent calls)
• User activates with "lean", "lean-output", or specifies a mode (lite/full/ultra)
• Speed matters and shorter responses preserve full accuracy

Don't use for:

• Explanations to non-technical users where full prose aids comprehension
• Documentation or user-facing content requiring complete sentences
• Creative writing where style is the requirement