Use the guide to communicate placement
The manual canvas becomes a real third reference image. It guides composition; it does not guarantee exact final product position or scale.
AI AVATAR / SHOTTIDE
A product avatar, built around your product.
Bring a product image and a presenter portrait. Guide the placement, review the generated composition, then add a voice when you are ready.
Choose the right starting point
A product avatar combines a presenter with a specific product reference. The first output is an image. You can keep that composition as a product visual or send it to the talking-avatar workflow for a separate video generation.
How to create a product avatar
- 01
Supply two distinct images
Upload the product and the presenter separately. Describe the product facts that must stay visible, including shape, packaging and color.
- 02
Choose automatic or manual placement
Let the model choose a placement, or move, resize and rotate the product on a layout guide. Select hold, point or demonstrate and describe the scene.
- 03
Review before adding speech
Check the face, fingers and product details in the generated image. Reuse the saved settings for a revision, or send the selected result to the talking-avatar tool.
What to check before you publish
Check the details that sell the product
Inspect logos, labels, dimensions and hand contact. Regenerate or edit an inaccurate result before using it on a product page or in an ad.
Inputs, limits and credit costs
- Inputs
- Two different owned images; each up to 5 MB and 16 MP.
- Composition
- Qwen Image Edit, 1 credit per image generation.
- Aspect ratios
- Six project ratios. Reference-based 21:9 uses a 16:9 source composition; final framing requires cropping or fitting.
- Video
- Narration and talking video are separate, explicitly requested steps.
Frequently asked questions
How does manual placement affect the result?
The guide shows the requested product position, size and rotation. The model receives it alongside the product and presenter references. Final pixels can differ.
Can the presenter demonstrate the product?
Choose hold, point or demonstrate-use as the image direction. Review the output carefully; a requested action is not a guarantee of physically accurate contact.
Does Generate also create a talking video?
The product-avatar action creates one composition image. After reviewing it, use the talking-video handoff and approve the separate speech and video costs.
Can I reuse a successful composition?
Yes. Saved jobs retain references, scene, action, framing and placement settings. Reuse those settings or use the finished image directly.