Hugging Face demonstrated an innovative approach to agentic AI by showcasing how an autonomous agent successfully constructed a three-dimensional virtual gallery of Paris by chaining together two separate Hugging Face Spaces. The project highlights the emerging capability of AI agents to orchestrate multiple specialized tools and services in sequence, autonomously managing workflows that would traditionally require manual coordination. This development illustrates a significant step forward in agent architecture and the practical integration of modular AI components. The experiment leverages Hugging Face Spaces—the platform's hosted environment for running machine learning applications—as interconnected building blocks for agent tasks. By chaining these spaces together, the agent demonstrated decision-making capabilities and task decomposition, breaking down the complex goal of creating a 3D gallery into sequential subtasks. The Paris gallery project serves as a compelling proof-of-concept for how agents can act as intelligent orchestrators, combining multiple AI models and services to achieve ambitious creative and technical objectives without human intervention between steps.