Clean raw HTML into LLM-ready text before agents spend tokens.
Free
Scanned where there is a package or repository to read, and every release diffed against the tool surface we already hold.
Clean raw HTML into LLM-ready text before agents spend tokens.