In this episode of The Build, Cameron Rohn and Tom Spencer dive into the evolving landscape of agents as MCP tools, unpacking how AI-driven applications are redefining what it means to build in public. They begin by examining the rise of Langgraph Agent and Langgraph Cloud, highlighting how these products enable rapid prototyping and scalable deployment of AI agents. The conversation then shifts to the technical architecture underpinning these systems, with a spotlight on the A2A design pattern and the MCP Tool Framework. By referencing integrations with platforms like Vercel and Supabase, the hosts illustrate strategies for seamless API integration, robust memory systems, and streamlined authentication workflows. They explore the practicalities of designing agent-based applications, discussing the benefits and challenges of leveraging Tool Swarm and SQL Agent for complex orchestration and data management. Throughout, Cameron and Tom emphasize the value of building in public, sharing real-world experiments, and iterating openly with developer communities. Entrepreneurship insights surface as they consider custom agent services and the emerging potential of an agent marketplace, encouraging founders to capitalize on open source collaboration and rapid feedback cycles. The episode concludes with a forward-looking message: as AI agents and MCP tools mature, developers and entrepreneurs who embrace transparent experimentation and robust technical architecture will be best positioned to shape the future of intelligent applications.