How to Turn a Photo into a 9:16 AI Video

Rules and credit examples checked October 29, 2026. Confirm the current quote before generating.

Choose Image to video, upload your photo, select a model that supports 9:16, describe one small movement, check the credit quote, and generate. Download the completed file and verify its actual dimensions before publishing. This produces a short moving clip; captions, accurate product labels and a finished advertisement may need a separate editor.

Step by step: from a photo to a vertical clip

1. Use a photo you have permission to upload. Keep the main subject sharp and away from the edges; a portrait source makes vertical framing easier. Do not assume a landscape image will preserve everything when reframed.
2. Open the generator, select Image to video, upload the image, and choose Wan 3.0 → 9:16. Check the available image-format and size limits in the uploader.
3. For a small motion study, choose 2s / 480p / audio on (53 credits). For a longer HD attempt, 5s / 720p / audio on costs 263 credits. These are different outputs, not a quality ranking.
4. Write one subject action and one camera direction. Preserve a copy of the original image and prompt.
5. Confirm the quote and available balance, submit once, and wait for the task status. Inspect and download a successful result before starting another attempt.

Prepare the photo before spending credits

Use a sharp source image with one clear subject. Leave space above and below the subject for a vertical composition, and avoid placing essential labels near the edges. Keep the original file so you can compare it with the output. The uploader’s current format, file-size and reference-image limits take precedence over general advice.

For a product shot, decide what must remain accurate: shape, color, cap, logo and label text. Requesting preservation does not guarantee it. If those details change, do not treat the result as an accurate product demonstration.

A product-photo prompt you can adapt

The bottle stays upright and still. The camera slowly pushes closer. Preserve the visible bottle silhouette, cap and color. Keep the background stable. No new objects, hands or on-screen text.

Replace “bottle” with your subject. Preservation instructions are requests, not guarantees: check the output against your original. This is a suggested prompt, not a newly tested product-video result.

What if the shape, label or framing changes?

• Distorted shape or text: reduce motion and keep the camera steady. Do not use an inaccurate product representation in an ad; add verified logos, prices and captions afterward in an editor.
• A cropped subject: leave more space around the subject in the source and inspect the first and last frames. Selecting 9:16 is not proof that every element survived the crop.
• An overly static clip: ask for one gentle movement, then review before committing to a longer request. Each completed revision can cost more credits.
• A pending or failed task: check task history before resubmitting. A confirmed failed or canceled charged task returns credits; an unwanted completed result is a creative rejection, not a technical failure.

Choose a vertical composition

Open the generator and choose Text to video or Image to video. Select a model, then choose 9:16 when the model supports it. The ordinary generator currently starts with Wan 3.0 at 9:16, two seconds and 480p. A saved draft or a specialized effect page may load different settings.

For text generation, describe one achievable action, the framing and the camera movement. A short prompt such as “A paper boat moves slowly across a calm pond, centered vertical composition, steady camera, soft morning light” gives a clearer starting point than several competing camera instructions.

For image generation, upload the intended starting image and explain what should move. Keep important subjects clear of the frame edges. Model capabilities and accepted reference inputs differ, so use the controls available for the selected model.

Check the quote before generating

The credit figure beside Generate video reflects the current request. In the current configuration, Wan 3.0 at two seconds and 480p costs 53 credits; five seconds and 720p costs 263. Changing settings changes the budget.

Sign in and ensure your balance covers the request. Daily check-in rewards accumulate over time; a new account does not immediately receive the full seven-day reward.

Inspect the downloaded file

Aspect ratio and resolution are different. A 9:16 composition describes shape, not a promise of 1080 × 1920 output. After a successful task, download the file and inspect its actual dimensions, length, audio and visible marks.

Before publishing, check the target platform's current upload requirements. Use applicable product and model terms to assess usage rights; neither spending credits nor selecting 9:16 establishes commercial permission by itself.

Keep a useful record

Save the original input, prompt, model, settings, credit quote, date and downloaded file. Record actual dimensions, duration and any visible marks. Describe editing or upscaling separately. This guide explains the workflow; it does not establish output quality or watermark status for every model.

Inspect archived portrait outputs

The first-frame guide includes two archived Seedance 2.0 outputs requested at 9:16. Both stored files measure 720 × 1280 and approximately 12.097 seconds, with audio. Their historical request cost was 1,912 credits each; they are not examples of the 53-credit Wan configuration above. The guide distinguishes verified file properties from missing input and retry records.

Inspect the archived result yourself

This is an archived Seedance 2.0 multi-reference output from October 16, 2026, requested at 12 seconds, 720p, 9:16 with audio. File inspection on October 30 found 720 × 1280 pixels, approximately 12.097 seconds and an audio stream. The archived request lists 1,912 credits; this is not a 53-credit Wan sample or a free-generation demonstration.

The original input bundle, exact complete prompt, generation waiting time and all-attempt denominator are not available here. This output cannot establish product accuracy, success rate, a model ranking or a watermark policy. Compare the visible first and last frames; see the archive notes and second output for the evidence limits.

Can I try this without paying?

The daily check-in guide explains how seven consecutive claims provide 64 credits, enough for one 53-credit attempt from zero. It does not cover a 263-credit request. For repeat attempts, see the one-time credit budget.