Wan2.7-Video

What is Wan2.7-Video?

Wan2.7-Video is a series of video generation models officially released by Alibaba Tongyi Labs on April 3, 2026, covering text-generated video, graph-generated video, reference-generated video andVideo EditingFour major models. The model supports text, image, video, audio full-modal input, focuses on the whole chain of video creation, covering the process of generation, editing, reproduction, remodeling, driving, continuation, reference, etc., which is said to be more controllable, more omnipotent, and “able to direct and excel in acting”.

Main functions of Wan2.7-Video

1. Full Modal Input and Unified Authoring System
   • It supports full-modal input of text, image, video, and audio, realizing a complete workflow from zero creation, screen control, interpretive reproduction to fine editing.
   • The entire series supports 720P and 1080P resolution output, and the video duration can be specified in any range from 2 to 15 seconds.
2. Powerful video editing capabilities
   • In a single sentence, you can make partial or overall changes to the video, such as adding, deleting and replacing elements, switching scene environments (e.g., sunny to rainy, summer to winter), and changing the overall visual style (e.g., realistic to watercolor, live-action to claymation).
   • Support commands to add and delete elements, replace objects, modify object properties, etc., such as “delete the train in the video”, “replace the film with a plate”, you can also modify the building color.
   • It supports precise addition with reference to the image content, and the edited area can naturally blend with the original video in terms of light and material.
3. Character and plot control
   • Character behaviors, expressions and emotions can be adjusted, and character lines can be replaced and automatically matched to lip-sync and timbre.
   • The shooting style can be redefined, including parameters such as camera position, viewpoint, field of view and focal length, to present a very different viewing experience in the same time and space.
   • Supporting intelligent script creation and split-screen scheduling, you can orchestrate the rhythm of the plot, design the language of the camera and arrange scene transitions with just a short text description.
4. Video Continuation and End-Frame Control
   • Introducing the joint mechanism of video continuation and end frame control, creators can specify the end frame while continuing the video, which not only ensures the continuity of the dynamics, but also realizes the precise control of the picture structure, effectively eliminating the “brake feeling” commonly found in the traditional first and last frame solutions.
5. Visual style and sound control
   • Thousands of style combinations can be extended to cover 2D cartoon, 3D animation, ink, clay, celluloid and other visual languages.
   • Sound realism, naturalness, musical expression and audio-visual synchronization quality are further improved, thanks to the Tongyi Labs voice team's technical accumulation in audio pre-training, audio data, and audio understanding.
6. Reference Raw Video with Action Replicas
   • It supports fast reproduction of movements, movement and special effects, and allows users to input image, video and audio multimodality, reference lock the appearance and tone of the character, and accurately migrate the character's movements, so that even large and complex movements can be stably reproduced.
   • Supports up to 5 video subject references, the most in the industry.

Wan2.7-Video Usage Scenarios

1. film and television production
   • It is used in pre-production processes such as script creation, sub-scene scheduling, scene conversion, etc., to reduce production costs and improve production efficiency.
   • Used for post-editing, such as adding, deleting and replacing elements, switching scene environments, changing the overall visual style, etc., to enhance the quality of the movie.
2. Creative Advertising
   • Quickly generate advertising videos to meet diverse creative needs.
   • Secondary creation of existing advertisement video to enhance the advertisement effect.
3. content creation
   • Provide self-media and short video creators with convenient video generation and editing tools to lower the threshold of creation.
   • Support personalized video creation to meet the creative needs of different users.
4. Education
   • Produce instructional videos, such as animated demonstrations and experimental simulations, to improve teaching effectiveness.
   • Edit and optimize existing teaching videos to improve teaching quality.

How to use Wan2.7-Video?

1. Visit the official website
   • Users can access the Manphase official website,wan.video website and AliCloud Hundred Refined Experience Wan2.7-Video.
   • Wan2.7-Video will also be available on the Thousand Questions App soon, so users can look forward to subsequent updates.
2. Select Creation Mode
   • Select text-born video, graph-born video, reference-born video or video editing mode as required.
3. Input creation instructions
   • Enter creative commands such as text, images, video or audio in the selected mode.
   • For the video editing mode, local adjustments to the video screen can be made by commands, such as adding, deleting and replacing elements.
4. Generate and download videos
   • Click the Generate button and wait for the model to process and generate the video.
   • Preview the generated video and download it to your local device when you are satisfied.

Recommended Reasons

1. Total creative control
   • Wan2.7-Video expands AI capabilities from single material generation to the entire creation chain, moving from “acting” to “guiding” to provide users with full creative control.
2. Powerful video editing capabilities
   • In one sentence, you can make partial or whole changes to the video to meet diverse editing needs.
   • Support intelligent script creation and sub-scene scheduling, reduce the threshold of creation and improve the efficiency of creation.
3. Rich visual style and sound control
   • Thousands of style combinations can be extended to cover a variety of visual languages to meet the aesthetic needs of different users.
   • Sound realism, naturalness, music expression and audio/video synchronization quality are all further improved to enhance the overall quality of the video.
4. Flexible video continuation and trailing frame control
   • Introducing the joint mechanism of video continuation and end-frame control, effectively eliminating the common “brake feeling” in the traditional first and last frame scheme, and improving the dynamic continuity and structural controllability of the video.
5. Multimodal referencing and motion replication
   • Supports image, video and audio multimodal referencing for accurate migration of character movement and appearance timbre.
   • Supports up to 5 video subject references, the most in the industry, to meet the needs of creating complex scenes.