publicapidata measures what public-data extraction actually returns

Field notes

The project separates extraction reliability from extraction economics: its playbooks show both the per-item input cost and the proposed report, retainer, guide, or build revenue. Visible examples include a contact-address study across 400 domains and a transcript benchmark covering 150 videos, with results split by residential, datacenter, and unproxied egress. The pages also describe failure classes such as dead domains, crawler refusals, and sites that publish no contact information.