wan2.6-t2v is the text-to-video model in the Alibaba Wan 2.6 series, designed for production from text ideas to narrative short films. It organizes visuals through intelligent shot scheduling, emphasizing consistency of subjects, scenes, and atmosphere across different shots. It natively supports up to 15 seconds and includes audio input and output capabilities. It is suitable for product stories, social content, and short-film concept previews, especially during creative stages when image or video assets have not yet been prepared.
Submit requests to the public API at api.acedata.cloud using the documented parameters, then use the results in your application.
Input parameters and result formats vary by service. Use the public API for this model and follow its guide for generation, task retrieval and editing operations.
Duration and resolution are described according to this model's native specifications; the audio toggle, shot options, and task delivery are explained according to this platform's invocation method.
wan2.6-t2v is suitable not only for describing a dynamic moment, but also for organizing continuous plots into multi-shot short films. When creating, clearly write the sequence of events by opening, main action, and ending, then add shot scale and atmosphere. Intelligent shot scheduling provides organizational capability for short stories, making it suitable for expressing a clear visual theme within a limited duration.
The focus of multi-shot creation is to connect characters, scenes, and atmosphere, rather than stitching together unrelated visuals. This model emphasizes consistency across these elements. Prompts can fix the subject's appearance, clothing, location, and lighting while varying only actions and viewing angles, making it easier to evaluate whether the story is coherent.
Native capabilities include audio input and audio output, making it suitable to treat sound as part of short-film design rather than considering only silent visuals. When invoked on this platform, audio is disabled by default and must be actively enabled for sound-enabled tasks. When writing scripts, describe the intended sound at the same time, then check whether the audiovisual expression matches the creative intent after completion.
Enter a textual description of the product use scenario, subject actions, and ending shot to generate a short video that showcases the creative idea. For example, design a departure, use, and arrival storyline around travel products for internal discussion of communication direction. If existing product images must be reproduced exactly, use an image-starting creation model instead.
Write a short-video idea as a brief shooting outline, describing the opening hook, main actions, and ending state, then generate a dynamic draft for review. It is suitable for comparing different story openings or shot arrangements. Change only one creative focus at a time to make it easier to determine which expressions are worth continuing to produce and edit.
Enter character settings, scene atmosphere, and the sequence of actions to turn a text storyboard into dynamic previsualization material. Directors, designers, or content teams can use it to discuss visual rhythm and shot transitions before deciding on subsequent shooting plans. The deliverable is a short-video previsualization and should not be treated as a final cinematography plan locked frame by frame.
When you only have a script, scene concept, or creative direction, wan2.6-t2v is the direct choice. If you already have an image and want to use it as the opening frame, choose wan2.6-i2v; if you want to extract a character's appearance from an existing video and bring it into a new scene, consider wan2.6-r2v. Prioritize choosing the model based on the creative starting point rather than sending all materials to text-to-video processing.
For one subject action or a complete presentation process, use single first to organize a single-shot expression; for a short story with clear plot turns, try multi. wan2.6-t2v has a native duration limit of 15 seconds and is suitable for focusing on one theme. For longer narratives, split them into segments and arrange post-production transitions rather than stacking too many people, locations, and events into a short video.
In prompt, clearly describe the subject, environment, action sequence, shots, and sound goals; this model starts from text, so do not use a first-frame image as a substitute for description.
Specify model=wan2.6-t2v and action=text2video for /wan/videos, starting with 5 seconds and 720P; Wan 2.6 does not use Wan 3's 30 seconds, automatic duration, or all-type media parameters.
Use async=true to save task_id, query /wan/tasks or receive a callback; after completion, retrieve video_url, check the subject, actions, and sound, then proceed to editing.
Early morning in a café, a barista grinds beans, pours water, and hands over a cup of coffee; the scene presents three distinct actions in sequence, keeping the person and in-store lighting consistent, with sounds of the coffee machine and soft clinking of cups and saucers.
Use text rather than reference images; first check the action sequence and shot continuity, then preview the audio; complex processes are best produced as storyboards, rather than assigning all multiple reference assets in a series to T2V.
This model is designed for text-to-video generation. If creation requires an existing image as the starting frame, choose wan2.6-i2v; if you need to continue a character's appearance from a video, choose wan2.6-r2v. Selecting the appropriate model based on the source material helps clarify how the visuals are controlled.
wan2.6-t2v has a native maximum duration of 15 seconds, with 720P and 1080P resolution options. It is suitable for creating short videos focused on a single subject. For longer stories, it is recommended to split them into multiple clips, then connect them through editing while maintaining the same subject and scene settings for each segment.
Write the prompt as a brief shooting outline: first describe the subject and scene, then describe the opening, main action, and ending in order. Keep the character's appearance, environment, and atmosphere consistent, and avoid redefining the subject for each shot. When multiple shots are needed, you can select shot_type as multi and review the generated continuity.
audio defaults to false in requests on this platform, so audio creation should explicitly set audio to true. This model natively supports audio input and output, but you should still check how the sound matches the visuals. If dialogue, music, or ambient sound has strict production requirements, further editing can be done after generation.
Submit a request to POST /wan/videos with model set to wan2.6-t2v, use prompt to describe the video content, and use text2video for text-to-video generation. After setting async to true, obtain task_id, then query the final status through /wan/tasks; successful results may provide information such as the video URL, dimensions, and thumbnail.
Choose a development approach or application suited to this model and follow its setup guide.
↗Python SDK Integration Guide
SDK installation, authentication and examples↗TypeScript SDK Integration Guide
SDK installation, authentication and examples
↗Pricing
$1.28–$4.42· per generation
You pay$13,584for 142,800 credits · $0.0951 / credits46% OFF
You're billed in credits. Pick a top-up tier to see its per-credit rate — the larger the top-up, the cheaper each credit.
This is only a basic call example. See the full docs for more parameters and advanced usage.
View Docs