> ## Documentation Index > Fetch the complete documentation index at: https://platform.vidu.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # Platform Introduction The Vidu Q series is China's first long-duration, high-consistency, high-dynamics video generation model, released by Shengshu Technology. Vidu holds a leading advantage in semantic understanding, inference speed, and motion dynamics, and has launched the world's first "multi-subject reference" feature, breaking through the challenge of consistency generation in video models and ushering in the era of visual context. The Vidu S series is built on a globally leading streaming inference video model released by Shengshu Technology, providing enterprise (B2B) customers with a new generation of near-real-time, interactive, bidirectionally-aware digital human services. It helps customers quickly deploy video AI characters that can "interact, perform, perceive, and provide long-term companionship" in scenarios such as social media, e-commerce, gaming, and education. By building a large-scale inference platform and optimizing the inference pipeline, the Vidu open platform transforms model capabilities into feature-rich, easy-to-use APIs. Enterprises and developers can easily convert multimodal inputs into high-quality video content, bringing richer interactive experiences to users across fields such as interactive entertainment, animated short dramas, e-commerce advertising, and cultural tourism/education. ## Video Generation Model Q Series: Reference-to-Video: Input a reference image and text description to generate a video. Currently supports referencing people, objects, and other types of subjects; Image-to-Video: Input a first-frame image, or first-and-last-frame images, along with the corresponding text description, to generate a video; Text-to-Video: Input a piece of text to generate a video; currently supports a general style as well as a specially optimized anime style. ### Use Cases Social media interactive features: including two-person hugs, animating old photos, personalized stickers, and more; Short content production for animation, games, etc.: generating video clips from a story script and editing them into a finished piece; Production of daily interactive promotional content for virtual IPs: production of IP promotional shorts and interactive video generation; E-commerce advertising production: generating empty-scene advertising footage, generating special-effect shots from static images, etc. ### Advantages Large motion dynamics and imaginative output: generated videos feature significant motion, without the sluggishness of "AI PowerPoint," making them well-suited for imaginative visuals or special-effect scenarios; Strong semantic understanding: accurately follows input prompt instructions; in particular, the image-to-video capability can follow prompts to generate a variety of scenes, making iteration ("card drawing") more efficient; Outstanding anime style effects: text-to-video generation produces better anime-style video results, and using anime images for image-to-video generation delivers high style stability with rarely any style drift; Stable subject reference results: overcomes the consistency problem that is difficult to solve at the image level, achieving highly consistent subjects. There is no longer a need to produce a first-frame image to generate a video with a consistent subject, making creation simpler. ## Real-Time Interactive Model S Series: Vidu S2-Avatar: supports real-time voice interaction and complex motion control, and can combine reference images to complete object interaction, costume changes, and background switching; Vidu S2-Editing: enables continuous real-time editing of an input video stream based on reference images for style, costume, subject, and background. ### Use Cases AI companionship, virtual idols, film/game interaction, training and instruction, AI customer service, e-commerce livestreaming ### Advantages Commercial-grade product: the first commercial-grade interactive digital human capable of interacting, performing, and bidirectional perception; Unlimited interaction duration: the world's first generative video foundation model technology supporting unlimited interaction duration, generating from 1 minute to 2 hours without quality loss; Excellent instruction response efficiency: supports instruction following with strong semantic understanding and the industry's fastest inference speed, enabling near-real-time cross-screen interaction; Warm interactive experience: supports multiple customizable personas and appearances, including real people, anime characters, and cute pets, with short-term memory, providing a personalized and warm interactive experience; Multimodal interaction formats: supports voice, text, video, and other interaction methods, accurately capturing factors such as user appearance and emotion; Excellent visual quality: high-resolution real-time interactive generation. For more features and model capabilities, refer to the [Feature List](https://platform.vidu.cn/docs/function-list) and [Model Map](https://platform.vidu.cn/docs/model-map)