Wan 2.6 AI Video Generator for Text, Images, and Source Clips

Start with text alone, exactly one image, or one to three videos. Do not combine images and videos, and do not upload audio. Every mode needs a written prompt: after leading and trailing spaces are removed, text-only prompts must contain 1–5,000 characters, while prompts with an image or videos must contain 2–5,000 characters.

For image input, use one JPEG, PNG, or WebP image with both width and height at least 256 pixels. For video input, use one to three different MP4 or MOV files. Each image or video must be no larger than 10 MiB (10,485,760 bytes). Sign in before uploading and use the uploader with Wan 2.6 selected. If you submit through the API, keep the file metadata from the original Files response with each reference.

Choose 720p or 1080p. Text and image modes offer 5, 10, or 15 seconds; video mode offers 5 or 10 seconds. The starting settings are 1080p, 5 seconds, shot mode Off, and content filter On. You can turn either control on or off; choosing Off is honored for both.

Submitting a video reserves the initial estimate: at 720p, 5 seconds reserves 70 credits, 10 seconds reserves 140, and 15 seconds reserves 210; at 1080p, 5 seconds reserves 105 credits, 10 seconds reserves 210, and 15 seconds reserves 315. The 15-second options apply only to text and image inputs; video inputs allow 5 or 10 seconds. Valid reported usage is rounded up to a whole credit and can be settled only within the reserved amount. If it is lower, the difference is returned. Missing or invalid usage, or an amount above the reservation, needs a billing review. A finished video does not by itself mean the charge is final.

Name the subject, the change you want to explore, and any camera direction or visual treatment that matters. For example, try a different setting around a product or ask for a new lighting direction. Compare the generated clip with your source when deciding what to try next.

Try a simple scene in text, add a product still to explore a motion idea, or upload existing footage to test a different visual treatment. Pick one input route for the run and write a prompt describing what you want to see.

Set the resolution to 720p or 1080p and choose 5 or 10 seconds in any mode. Text-only and image inputs also offer 15 seconds. Shot mode and the content filter each have an on/off control. The form starts at 1080p and 5 seconds, with shot mode Off and the content filter On; either can be switched off.

Sign in, select Wan 2.6, and upload either one JPEG, PNG, or WebP image or one to three different MP4 or MOV videos. An image must be at least 256 pixels wide and 256 pixels high. Every file has a 10 MiB (10,485,760-byte) limit. Do not mix image and video references or add audio. A prompt is still required with your files. For API submissions, pass along each file's original upload-response metadata.

You can browse this page before signing in, but uploading and submitting a run require an account. Submission reserves an initial estimate, not a fixed final charge. The 720p estimates are 70 credits for 5 seconds, 140 for 10 seconds, and 210 for 15 seconds. At 1080p they are 105 credits for 5 seconds, 210 for 10 seconds, and 315 for 15 seconds. Only text and image inputs offer 15 seconds; video inputs are limited to 5 or 10 seconds. Final billing rounds valid reported usage up to a whole credit and settles it only when it is within the reservation, returning any difference if it is lower. Missing or invalid usage, or usage above the reserved amount, must be reviewed. A completed video can still be awaiting final billing.

Open the Studio to work with the generation form, adjust the input and settings for another run, or follow the links to your generation history. You can also start with a test on this page while deciding which input route suits your idea.