AI creation: Turn a sentence into a video

Sign in and open Workspace → AI creation to generate a vertical short video from a visual description. The price is shown before charging your wallet, with automatic refunds on failure. Other tools include image-to-video, text-to-image, image editing, video extension, titles and topics, and video upscaling. Video analysis and Talking avatars have separate entries; the Voice library is under Talking avatars. Desktop also supports BYOK and long batch tasks.

After signing in, hover over AI creation in the left navigation to open a tool directly. On mobile, expand its submenu. Empty result panels explain the steps; content marked as a demonstration illustrates the workflow.

Hovering over AI creation opens tools including text-to-video, image-to-video, text-to-image, and product photo processing — Hover and select a tool. Chinese web interface preview with demo account and quota values; the live page is authoritative. Click to enlarge.

Steps: Text-to-video example

Open Text to videoSign inWeb app, then open Workspace → AI creation → Text to video.
Write a visual promptDescribe the subject, setting, and camera motion, or use Try example for a sample prompt. Exclude blur, watermarks, and other unwanted elements in the negative prompt field. The web app accepts one prompt per submission.
Choose a model and sizeChoose a model and frame size, such as vertical 9:16. Available options depend on the model.
Generate and saveClick Start generating, review the quote, and confirm. Preview and download results on the right or check Recent generations. Save temporary video links promptly.

Navigation: Video analysis and Talking avatars have separate entries. Voices are in the Talking avatars submenu; video generation and product photo tools are under AI creation.

Available AI creation tools

Text-to-video interface: prompts, model, and size on the left; a three-step guide and result canvas on the right — Set parameters on the left; see instructions and then results on the right. This Chinese interface uses a sample prompt without submitting a job. Account and quota values are demonstrations. Click to enlarge.

Product photo processing: Ten operations, including watermark removal, white backgrounds, outpainting, and upscaling, plus AI product descriptions (see Product photo guide）。
Text to video: Generate a vertical short video from a visual description.
Image to video: Upload an image and describe how to animate it into a video.
Text to image: Generate images from prompts with selectable count and aspect ratio.
Image editing: Edit selected areas or describe edits to an original image.
Video extension: Extend an existing short video using a prompt.
Video analysis: Extract dialogue, prompts, and shot references as needed.
Titles & topics: Generate attention-grabbing titles and topic suggestions from a theme.
Voice library: Record or upload your voice to clone it for voiceover and avatars.
Talking avatars: Generate video with audio from a script and optional reference voice and avatar. Reference fidelity may vary (see Talking avatar guide）。
Video upscaling: Upscale low-resolution videos to 1440p using AI (see Video upscaling guide）。

Billing: The web app uses platform billing: official keys, prices shown before each job, charges from your wallet, and automatic refunds on failure. Desktop app also supports BYOK: connect directly with your Alibaba Cloud Model Studio key and bill your cloud account, without platform charges. Desktop is also suited to batch and long tasks.

Frequently asked questions

How long does generation take, and how long is the video?

Wait times depend on the model, input, and queue. Output duration depends on the selected model. Check progress on the right and download temporary outputs promptly.

Will failed generation be charged?

With platform billing, reserved funds return to your wallet automatically on failure. View transactions in Account → Balance & billing.

Is this on the web or desktop?

On the web, open Workspace → AI creation and pay from your wallet by usage. Windows and Mac additionally support BYOK and long batch tasks. Both share your account and membership benefits.

Can I use a custom model?

Platform billing supports only listed models with available prices. For custom models, use BYOK in the desktop app with your own key.

→ Transcribe uploaded videos → Product photo processing → Talking avatars → Video upscaling → Desktop app settings ← Back to Help