Updated 2026-10-08
Hotel Lobby AI Character Swap
Hotel Lobby AI character swap uses two photos as references for replacing characters in an orange-booth performance. The reference video supplies the scene, movement and soundtrack. A photo changes appearance, not the singer’s voice.
New generation and credit purchases are coming soon. This guide explains the workflow being prepared; it is not an invitation to submit a paid task today. Check Pricing for the current status.
Decide who takes each role
- Watch the template preview, including its sound. Identify the lead performance and any accompanying role.
- Choose one clear subject per photo. Photo 1 maps to the character starting on the viewer’s left; Photo 2 maps to the character starting on the right.
- Keep a record of that assignment. A character moving across the frame should not cause the identities to switch.
- When generation becomes available, compare the opening, movement and ending with both input photos before judging the result.
Use the photo preparation guide to choose useful references. Visible facial detail helps recognition; including clothing gives the model more information about appearance. Neither guarantees an exact reproduction.
Appearance, performance and voice are separate
This workflow creates replacement imagery; it is not a guaranteed pixel-perfect face overlay. Hair, clothing, hands and facial features may change during motion. An output can finish successfully yet still need a visual review.
Two people in the scene do not automatically mean two vocal parts. The actual reference controls the intended choreography. Replacing an accompanying character with a woman does not add a female vocal, and replacing the lead does not change the existing recorded voice. Lip movements can also diverge from the retained audio.
Check the evidence
Read the example and preview guide to distinguish a source preview from a personalized result. If a future result has the wrong roles or drifting features, use character-swap troubleshooting. The guides index connects preparation, review and finishing steps.
FAQ
Does a female photo produce a female voice?
No. Character replacement changes the visual reference. The finished clip retains the template soundtrack; the uploaded photo does not select or clone a voice.
Do both characters have to rap?
No. The reference determines the performance. In a single-lead arrangement, one character performs while the other accompanies. Watch the actual preview instead of assuming a duet.
Can I swap only one person?
The configured workflow uses two separate character photos. It does not offer a dedicated one-person replacement mode. New generation is currently unavailable.
Can I generate a character swap now?
Not yet. New generation and credit purchases are coming soon. You can read the guides, view template previews and check the published credit costs.