Photo tips for recognizable AI rap performers
The source image determines how much identity, clothing and body detail the model can see. Use this checklist before generating a solo, duet or pet performance; prompt wording cannot recover detail missing from a photo.
Last updated:
1. Start with a sharp, unobstructed face
Choose even lighting and visible eyes, nose and mouth. Avoid motion blur, heavy beauty filters, dark sunglasses, a hand covering the face or a very small face in a crowd. A sharp ordinary photo is more useful than a dramatic but unreadable portrait. Identity preservation is requested, not guaranteed.
2. Choose framing for the movement you want
A headshot gives the model no evidence of trousers, shoes or full-body proportions. Use a head-to-toe image for dancing and footwork; a waist-up image suits restrained upper-body performance. Leave some space around the person. Changing the output to 9:16 does not recover body details cut off in the source.
3. Use the correct one-photo or two-photo layout
For a solo, use the single-photo option in a supported tool. For two independent people, choose separate photos and put one person in each slot. For an existing pair, use shared-photo mode with one image containing both. Avoid collages, overlapping portraits and photos containing extra bystanders.
4. Keep clothing and image quality readable
Use JPEG, PNG or WebP, up to 10 MB per image. Convert phone HEIC files before uploading. A screenshot of a screenshot often loses hair and fabric detail. Avoid pre-cutouts with rough edges; a clean original gives more context than a badly erased background. Visible logos or text may distort during animation.
5. Prepare pets and illustrated characters differently
For a pet, show the muzzle, ears, coat pattern and body without a cage or foreground obstruction; choose Pet Rap. For a cartoon or anime character, choose Character Stage and use an image you have rights to use. A character reference is not a promise to convert a real face into a specific drawing style.
6. Inspect these defects before spending again
A changed face calls for a clearer reference, fewer accessories obscuring it and simpler movement. Rough hair edges may come from the source or the generated background transition. Extra limbs can come from overlapping bodies or complex gestures. Try a simpler composition before repeating the identical request; no photo eliminates all model artifacts.
7. Check permission and minimize sensitive material
Use your own image or obtain the necessary permission from the subject; children’s photos require appropriate guardian permission. Do not upload identity documents or images containing private information. Your result is not permission to use another person’s likeness commercially. Review the privacy policy for source-photo handling and download finished videos within the 30-day retention period.
A brief you can adapt
Two friends, each in a separate clear photo. Keep the left performer’s red jacket and the right performer’s blue shirt. Ask for small shoulder bounces and hand gestures in a frontal two-shot. Use head-to-toe source photos if both pairs of shoes must remain visible.
Watch an existing example
Plan your first generation
Standard · Mini · 720p. These examples use the current site-credit calculation. Check the form estimate before submitting; available options and access are shown in the tool.
| Seconds | Site credits |
|---|---|
| 5 | 5 |
| 10 | 9 |
| 15 | 13 |
Account access, downloads and help
Sign in with Google or your verified email account, purchase a one-time credit pack, and check the credit cost before generating. No manual access application is required. Download completed MP4 videos within their 30-day storage period. Contact support if you need account or billing help.
[email protected]