Hugging Face has introduced vision support to its smolagents library, enhancing the capabilities of these AI agents. This update allows users to pass images at the initialization of the agent, enabling functionalities such as chatting with documents and automatically filling forms. Additionally, agents can now dynamically retrieve images, facilitating agentic web browsing with minimal changes to the code. This development is seen as a response to OpenAI's recent launch of its Operator feature. The smolagents library also includes secure code execution capabilities.