Vidu: Vidu Q1 Reference to Video API

Vidu Q1 Reference-to-Video is an early-generation reference image-driven video generation model developed by Shengshu Technology, part of the Vidu Q1 series. As the first iteration of Vidu's reference-to-video capability, Q1 laid the technical foundation for the subsequent Q2 series. The model supports video generation driven by reference images combined with text descriptions, with foundational capabilities in character appearance, style, and scene consistency. Key capabilities include: Reference Image-Driven: Uses one or multiple reference images as anchors to extract visual features of characters or scenes and transfer them into the generated video content. Text-Guided Generation: Provides basic control over video content, character actions, and camera movements through text prompts. Baseline Consistency: Maintains primary visual features from reference images throughout the generated video, suitable for creative scenarios with relatively relaxed consistency requirements. Lightweight Deployment: As a Q1-generation model, it consumes comparatively fewer computational resources, making it suitable for rapid prototyping or low-cost content production. Typical use cases include early AI video content exploration, low-cost short video production, and creative prototype validation.