Vidu: Vidu Q2 Turbo Multiframe API

Vidu Q2 Turbo Multiframe is the extreme-speed multi-frame-driven video generation model of the Q2 series developed by Shengshu Technology. As the speed-maximized version for multi-frame video generation in the Q2 series, the Turbo version takes maximizing generation speed as its core design objective, compressing the waiting time for multi-frame-driven video generation to the lowest level in the Q2 series through deep inference acceleration optimization. Unlike the start-end frame mode which only accepts two anchor images, the Multiframe mode supports users providing three or more frame images simultaneously as visual anchor points, combined with optional text descriptions, generating coherent video content that smoothly transitions between multiple visual keyframes at maximum speed, providing an extreme-speed solution for high-throughput application scenarios requiring precise multi-node control over video content. Key capabilities include: Extreme-Speed Multi-Frame Visual Anchor-Driven Generation: Accepts three or more user-specified frame images as multiple visual anchor points, outputting coherent video content that sequentially and smoothly transitions between all specified frames at the fastest generation speed in the Q2 series under a deep inference acceleration framework. The Turbo version places inference speed at the highest priority, drastically compressing the waiting time for multi-frame video generation, significantly improving platform throughput and user response experience, particularly suitable for real-time multi-frame animation scenarios with strict requirements for generation latency. Rapid Multi-Node Visual Path Construction: Under the Turbo extreme-speed inference framework, quickly identifies and processes the visual relationships and content evolution directions between multiple input frames, constructing the visual transition path from the first frame through intermediate anchor frames to the final frame with maximum efficiency, ensuring basic visual consistency and temporal correspondence of each anchor frame in the output video under the speed-priority premise. Efficient Multi-Segment Motion Trajectory Inference: Under the speed-priority inference framework, quickly infers motion trajectories between adjacent anchor frames in segments, generating video content with basic motion rationality and fluidity for each transition interval, maintaining an overall dynamic coherence level meeting basic application needs under the premise of extreme-speed generation. Text-Assisted Multi-Frame Guidance Support: Supports optional text prompt input, using text descriptions to provide basic auxiliary guidance on the overall transition style, motion tone, and scene atmosphere across multiple frames, providing basic text controllability for multi-frame video content within the speed-priority generation framework. Lowest Resource Consumption and Cost Advantage: The Turbo version has the lowest inference resource consumption and per-generation cost among the Q2 Multiframe series, suitable for large-scale concurrent multi-frame animation tasks, high-frequency batch multi-keyframe video content production, and commercial scenarios with limited cost budgets, representing the most economical choice in the Q2 series multi-frame video generation direction, effectively reducing the overall cost of large-batch multi-frame-driven video content production. Typical use cases include large-batch multi-keyframe video rapid generation, high-throughput multi-frame video content automated production pipelines, latency-sensitive real-time multi-frame animation content generation, low-cost multi-node video transition content rapid production, high-frequency rapid creative multi-frame video prototype validation, small-to-medium content creation projects with limited cost budgets, and content factory-type application scenarios requiring rapid output of large volumes of multi-frame transition videos.