Claygent Navigator is a 📝Claygent model inside 📝Clay that automates interactive web browsing — clicking buttons, filling forms, paginating, and pulling structured data from sites that don't expose clean APIs.
It exists because most outbound enrichment hits a wall at the page boundary: data lives behind login walls, multi-step search forms, paginated grids, or PDFs that text-only research agents can't read.
Built for 📝Go-To-Market (GTM) and sales-ops teams using Claygent to enrich accounts from sources that resist scraping — public registries, dashboards, gated databases, and multi-hop research flows.
Key Capabilities
- Interactive page handling — agents click buttons, fill out forms, toggle filters, scroll through long results, and paginate through gated UIs that resist plain scraping.
- Multi-format extraction — extracts structured data from web pages, dynamic tables, PDFs, and XML or CSV outputs, all in a single Claygent run with no separate parser.
- Replay visibility — every Navigator run includes a step-by-step Claygent Replay showing exactly what was clicked, filled, and checked along the way, ready for inspection.
- Multi-hop research — chains across multiple pages, dropdowns, and external sources in one continuous research run instead of stitching together a series of one-shot lookups.
- Operator-style agency — inspired by 📝OpenAI's Operator, Navigator acts on the page like an active teammate taking the next click instead of a passive scraper waiting for direction.
Getting Started
- Open a Clay table and add or open a Claygent column.
- In the Claygent settings, select Navigator as the model.
- Describe the task in plain language — what site, what to click, what to extract.
- Test on a few rows to confirm Navigator can reach the target data.
- Review the Claygent Replay for each row to verify steps before scaling up.