Metric: Token bloat

Token bloat estimates when boilerplate and non-content overwhelm the useful information on a page.

Interpretation

Higher token bloat usually means the primary content is a smaller portion of the overall HTML/DOM text. This can hurt extractability and can cause AI systems to focus on repetitive UI instead of your core message.

Common causes

Fixes

Where it appears

In JSON reports, inspect token_bloat_ratio (field name may evolve; treat the report JSON as the source of truth).