{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/0ce36502ca52452da8aa65cc589e004c\" frameborder=\"0\" width=\"3340\" height=\"2505\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":2505,"width":3340,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":2505,"thumbnail_width":3340,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/0ce36502ca52452da8aa65cc589e004c-6d5133c36cf07f57.gif","duration":405.294,"title":"C101 Bonus - Claygent","description":"This Loom highlights Claygent, emphasizing it as a key differentiator for accurate AI web scraping and personal data research. The speaker demonstrates building a Claygent that finds rank-ordered personal and professional information about an individual, prioritizing interviews, podcasts, blog posts, and personal media, then personality interests, then previous jobs, and finally other contextual details if needed. They run the prompt on Guillermo Rush and discuss key Claygent settings, including not always letting AI auto-pick the model and spending more credits on data points that matter, with personal data tested at 45 credits per use. The video also covers using account and business context plus adding tools, including custom MCP servers for integrating external systems."}