Web-agent is lowering the cost of web interaction from 'writing scrapers' to 'calling models.' By encapsulating search, scraping, and interaction logic, this open-source framework allows users to transform manual monitoring processes into fully automated backend tasks simply by integrating Anthropic or OpenAI models. The widespread adoption of this architecture could shrink the survival space for entry-level data analysts who rely on manual data scraping and web-based operations. The key question now is: when web interaction logic is fully driven by models, can traditional anti-scraping strategies still effectively block these highly intelligent automated agents?