@cinatra-ai/web-scrape-agent

Web Scrape Agent

Point this agent at a handful of web pages and a description of the shape of data you want — company directories, event speaker lists, batch pages, board rosters — and get back clean structured rows ready to feed into a list, a CRM, or a downstream agent. You stay in control of the shape; the agent does the visiting and extracting.

Works with

• The open web (via web search)

Capabilities

• Extract structured rows from a list of seed URLs
• Conform every row to a JSON Schema you supply
• Follow same-origin links to a bounded depth when you ask it to
• Deduplicate the same entity across multiple pages
• Attach a source URL to every extracted row
• Record per-URL failures so nothing is silently dropped