AI Product Talking Video Editor
Upload your original talking video and a garment or product image to request an outfit or product change along the original performance and timing. An optional presenter image guides identity.
ORIGINAL VIDEO × PRODUCT × PRESENTER
Restyle your original product presentation
Start with your original talking video and a clear garment or product image. Request a targeted outfit or product change while following the original action, camera work and timing. The six gallery videos illustrate presentation styles; upload the clip you want to edit, as gallery videos are not automatically used as source footage.
Creation examples
THREE SIMPLE STEPS
How to edit a product talking video
An original video and a garment or product image are required. Use the optional presenter image only when you want it to guide the person’s identity.
Upload your source video and product
Upload one 4–30 second MP4, up to 100 MiB, and one clear garment or product image. For apparel, show the whole garment; for a product replacement, show the complete target item.
Name the target and what to keep
Describe which person, garment or object to change, especially when several are visible. Specify any clothing, objects or identity to retain. A presenter reference is optional; the original person is retained when none is supplied.
Choose resolution and review the result
Editing follows the source duration and frame shape. Output defaults to 1080p; 720p is available. Check the estimated cost, then review faces, clothes, hands, product details, speech, timing and lip movements.
EDITING CONTROLS
Give each reference a clear role
The original video supplies the performance and timing. The product image supplies the replacement appearance, and the optional presenter image guides identity. Results still need review.
Garment or product appearance
Use the product image for cut, color, material, shape and visible branding. Name the garment or object it should replace. Reference backgrounds, unrelated props and a model in the product photo are not part of the replacement.
Optional presenter identity
A clear presenter image guides the requested face and appearance. Without it, the edit retains the original presenter. Identity consistency can vary, particularly with turns, occlusion or a small face in frame.
Original performance and speech timing
Request the same action order, camera work, spoken words, sound timing and lip movements while changing the specified appearance. Audio and mouth movements are generated and must be checked; this is not an exact audio-track pass-through.
EDIT DESCRIPTION
Describe a focused edit
For a try-on: “Change the main presenter’s shirt to the garment in the product image. Keep the original trousers, face, movements, camera work and spoken words.” For a product edit: “Replace the bottle in the presenter’s right hand with the product image; retain the person and all other objects.” If you also upload a presenter image, state whether to use that identity or keep the original person. New dialogue, when needed, should be short and placed in quotation marks; it takes precedence over the original words and requires new lip synchronization.
WAYS TO USE IT
Edit the presentation you already have
Use your own source performance and focus each version on a clear change.
Restyle an existing performance
Use the source video and garment image to request an outfit change along the original action and timing. Specify which pieces to replace and which clothes to retain.
Replace a featured product
Change a specified item using the product image. Request plausible hand contact, scale, perspective, lighting and shadows while retaining the surrounding performance.
Try a different presenter appearance
Add the existing optional presenter reference when you want a different identity while following the original performance. Review face consistency through speech, turns and occlusion before using the result.
EDIT YOUR FIRST VERSION
Start with the video you want to edit
Upload your original video and garment or product image. Name the target, state what to retain, and review the first result before refining the edit.
Frequently asked questions
What materials are required?
Upload one original MP4 video, 4–30 seconds and up to 100 MiB, plus one clear garment or product image. Images may be PNG, JPEG or WebP, up to 10 MiB each. The existing presenter-image field and edit description are optional.
How do I specify what to change?
Identify the person, garment or object by position or appearance, and name the clothes or objects to retain. Use @ material references when useful. The product image guides the replacement item; a model appearing in that image does not define presenter identity.
How long is the output, and which resolution can I choose?
Editing follows the original video’s duration and frame shape; this play does not provide a manual duration choice. A small timing difference is possible. Output defaults to 1080p, with 720p available. Resolution is not a guarantee of identical original detail.
Will the original audio and lip movements stay exactly the same?
The prompt asks to retain the original words, sound timing and lip movements unless you request a change. The model generates the result, so exact audio preservation or lip synchronization is not guaranteed. Short new dialogue in quotation marks replaces the original words and requires new mouth movements. Play back the complete result before use.
How is presenter identity chosen?
Without a presenter reference, the edit retains the source person. The optional presenter image guides a requested identity change. If you only want a clothing or product change, leave that field empty or explicitly ask to retain the original person. A reference helps guide identity but does not guarantee a face match.
How can I improve the replacement?
Use a clear, complete garment or product reference and a source clip where the target is visible. Begin with one focused edit. Review garment shape, logos, text, proportions, fingers, hand contact, shadows and continuity through occlusion.
Do gallery videos become my source footage?
No. The six videos illustrate presentation styles. Upload the original clip whose action, timing and dialogue you want to use for the edit.
How should I refine the next version?
Keep the same source video and product reference, then change one instruction at a time. State the target and retained details clearly. Compare identity, product appearance, performance timing and audio before using the result.