AI Video / Character Mapping

Hotel Lobby AI Video Generator

Video-to-Video

Replace people in the video while preserving the original shots, movement and sound.

Video preview

hotel_lobby.mp4

hotel_lobby.mp4 · MP4 / MOV / WEBM

Upload source video

hotel-lobby-source.mp4

Upload replacement references

Up to 4 images; hover to manage each reference

Add reference image(2/4)

My prompt (optional)

Use Video1 as the exact foundation. Preserve its scene, background, camera angles, camera movement, lighting, composition, pacing, character positions, actions, interactions, dialogue, voices, music, and sound effects.

Replace only the two original people. Replace the person on the left with Image1 and the person on the right with Image2. Preserve each reference subject's face, hair, colors, textures, body proportions, clothing identity, and distinguishing features consistently in every frame. Never swap, merge, duplicate, or redesign the subjects.

Lip synchronization is the highest priority. Each replacement subject must inherit the corresponding original speaker's timing, mouth shapes, facial expressions, conversational rhythm, pauses, and subtle facial motion frame by frame. A subject must stop speaking when the corresponding original person stops speaking.

Keep the original body motion, timing, spatial relationship, and interaction. Integrate both subjects naturally with realistic perspective, lighting, shadows, occlusion, and motion continuity. Preserve the original audio without creating a new voiceover.

Produce a photorealistic, cinematic result. Avoid identity drift, flicker, face or body deformation, unstable hands, incorrect lip sync, delayed mouth movement, character swapping, background changes, camera deviations, added text, logos, subtitles, watermarks, or visual artifacts.

ModelAspect ratio

1462 credits · 1080P · adaptive

Replace characters, not the whole video

Change identities without rebuilding the scene, camera, movement or sound.

Original

Mapped

Preserve scene

Preserve motion

Preserve audio

4 people detected

Map characters in three steps

Upload the source video

The system identifies each person in the clip.

Add a reference for each person

Match every detected person with a clear portrait.

Generate and download

Preview the finished video and download it in HD.

Original video 4 people detected

As many people as the video contains

The system detects each visible person so every identity can be mapped independently.

4 people detected

Change the cast, keep the video

Only the selected identities change. Everything that makes the source work stays intact.

Original scene and framing

Keep the location, lighting and composition.

Camera movement

Retain the source camera path and perspective.

Action and interaction

Preserve timing, gestures and interactions.

Original audio and rhythm

Keep speech, music and pacing aligned.

Made for every kind of cast

From real people to pets, map every role from the source video.

Generated example

Two cats can star in the same video

Upload the source video and the system identifies each role, then maps the pet references one by one. The result preserves the scene, camera, interaction and audio rhythm.

Role 1 · Left

Original reference

Role 2 · Right

Original reference

Preserve scenePreserve motionRole mappingPreserve audio

Try pet characters View full example

For pet references, choose clear front-facing, full-body images with distinctive clothing to keep each role easy to distinguish.

Frequently asked questions

How many people can I replace?

The interface supports multiple detected people. This Hotel Lobby template uses two mapped roles.

Does it keep the original sound?

Yes. Original dialogue, music and overall pacing are preserved.

What images work best?

Use a clear, front-facing or three-quarter portrait with one visible subject.

Will the result look exactly like the reference?

Results depend on pose, lighting and occlusion. Clear portraits usually produce stronger identity consistency.

Make the same video with a whole new cast

Upload two reference portraits and create your Hotel Lobby AI video.

Try pet characters