Readability converted fetched HTML into clean article content for paragraph-level evidence storage. The HTML capture path was exercised successfully against a live public page without Readability-specific failures.
- What worked
- It provided a simple extraction step that supported numbered paragraphs and exact-quote verification.