Connector Layer

Crawler Connector Abstraction

The public connector layer explains how web results, datasets, directories, and webhooks should normalize into a consistent intent-ready shape. It intentionally does not include private source seeds, production search templates, anti-blocking tactics, or abuse controls.

Connector Contract

Input: source, fetch_context, raw_payload
Output: normalized company record, contact hints, provenance, freshness, and extraction confidence
Downstream handoff: convert candidate records into supplier or buyer intent candidates for review

Why Keep It Abstract

A public connector SDK helps builders plug in their own sources. Production source lists, crawl cadence, anti-spam heuristics, and filtering rules remain private because they materially affect marketplace quality and operational safety.