OneSearch-VL: Unified Multimodal Deep Research Agent for Image and Video
OneSearch-VL is a unified agent designed for deep research across single images, multiple images, and videos. It uses a Visually Grounded Evidence Graph (VGEG) to link visual anchors to external facts and define the steps needed to form an answer. This structure allows the agent to handle complex reasoning tasks by organizing data construction, process supervision, and operation evaluation.