Start with the real product
Upload a well-lit photo, write its true benefits and a call to action, then pick a studio, everyday or creator-to-camera direction and whether the ad is spoken or music only.
PRODUCT PHOTO → 15-SECOND AD
Upload one product photo and write three short lines. MiniMax H3 shoots a 15-second vertical ad from that photo with the script spoken and sound in one pass. Add your logo and captions, then download the MP4.

Generation requires purchased model access and credits. The quote is shown before you generate.
View plans and credits ↗Upload a well-lit photo, write its true benefits and a call to action, then pick a studio, everyday or creator-to-camera direction and whether the ad is spoken or music only.
Edit the hook, the benefit and the CTA, or let AI draft them. Your photo becomes the reference image; the model shoots three beats around your words with voice and sound in one pass.
Review the generated ad, add captions, a CTA card and your logo, and download a 15-second vertical MP4. Re-export brand edits from the same video at no extra credits.
Earlier product-video tools rendered a still frame first and then animated it, which meant a slow push-in and a separate dubbing step. MiniMax H3 takes reference images directly: the product photo anchors identity in every shot while the prompt directs setting, camera and narration. This page writes that kind of prompt from your brief.
Prefer another model? The dropdown lists every model that can shoot a 15-second vertical clip from one photo. Reference models keep the product identical in each shot; first-frame models animate your photo, often for fewer credits.
Reference mode keeps shape, color and packaging consistent across shots instead of reinventing them. Still check fine print and logos before publishing; regenerate if a detail drifted.
The script is spoken by the model as the ad is generated, with a music bed and ambient sound. Choose music only, keep the generated voice, or replace it with a configured studio voice at export.
Export a 720 × 1280 MP4 in 9:16 for Reels, TikTok and Shorts. Captions, the CTA card and your logo are layered at export, so brand edits never cost generation credits.
Drafts and scripts are free. One generation quote covers the 15-second ad with voice and sound; regenerating uses it again. Captions, logo and CTA edits reuse the video. Optional studio narration has its own quote.
A purchased plan or credit pack unlocks the models used here. This page does not promise free video generation.
You can explore the tool, write a script and save a draft without spending credits. Generating the 15-second ad requires purchased model access and enough credits, and the exact quote is shown on the button before you start.
Yes. Upload a PNG, JPG or WebP and it becomes the reference image for every shot. Pick a studio showcase, an everyday setting or a creator speaking to camera. This version starts from a photo; it does not import a product URL.
The branded export is a 15-second, 720 × 1280 MP4 in 9:16 with your caption, logo, CTA and audio choices. The raw generated ad is also available to download. Publishing and campaign management are not included.
By default the model speaks your hook, benefit and CTA as it generates the ad, with music and ambient sound. Choose Music only for no speech. If a studio voice is configured, export can replace the generated audio with that narration for a separate quote.
Reference mode keeps the product far closer than a text prompt, but AI can still soften fine print or small logos. Review the generated ad and regenerate if a detail is wrong; regeneration uses the shown credits.
Captions, the CTA card, the color, the logo and the audio choice can be changed and exported again for free. Changing the spoken lines needs a regeneration, because the voice is part of the generated video.
Paid accounts export without the site watermark and can use their ads commercially under the commercial license. Review platform ad policies and your own claims before publishing.