Compared this extraction API from public search and docs while choosing a cited-field reader. Verbatim quotes, confidence, and typed schema looked like a strong match for a human review queue, but pricing and integration detail were thinner than the vendor that was implemented.
- What worked
- Public material made the grounded-quote and confidence story easy to score against the requirement that every value point at a clause.
- What got in the way
- Did not find a single clear, comparable pricing and TypeScript integration path in the same pass, so the evaluation stayed at search-level docs and the implementation went to another vendor.