How-to guide

How to make the First Time in a Fancy Cafe AI video

Two photos, one cafe POV remake. Keep the upload order clear.

Two adult photos15 seconds · vertical

Updated 2026-11-04

Upload two separate adult photos in order: photo one maps to screen-left and photo two maps to screen-right. Review the displayed price, generate, then check the video and audio before sharing.
  1. 01
    Add the screen-left photo

    Choose one adult whose image you have permission to use. This person takes the first role.

  2. 02
    Add the screen-right photo

    Choose the second adult. Keep this role paired with the second upload.

  3. 03
    Generate and review

    Check the price shown before generation. When ready, inspect character consistency, furniture interactions, duration and audio.

Before you start

Use one consenting adult per photo. A clear face, natural light and an unobstructed view help the model keep the two identities distinct.

  • Do not swap upload order after choosing the roles.
  • Use only photos you have the right to use.

What the reference video controls

The reference provides the cafe scene, action, camera, pacing and sound. The character photos provide visible identity and appearance for their assigned roles.

Make your own First Time in a Fancy Cafe AI video
Make my video

Back to the generator: AI video templates

AI-generated video
1 · Pick a template
+
More templates coming soon
2 · Add your photos
First Time in a Fancy Cafe AI
15s · 9:16 · 50 credits · Audio included
Photo tips · JPG / PNG / WEBP, up to 10 MB
  • ✓One person per photo, facing the camera
  • ✓Face clear and well lit — no sunglasses, hats or hands over it
  • ✓Show the outfit: a photo from the head to the waist or knees
  • ✓No mirror selfies or phone in hand — it can show up in your video

🎵 Audio is included in the video. Use Sound On or Sound Off to listen.

You see the price before you pay. Credits never expire. Pricing

Questions people ask

Which photo should go first?+

The first photo maps to the original person on screen-left. The second maps to the person on screen-right.

Does it copy the source video?+

The prompt asks the model to preserve the reference scene and timing while replacing the two people. Generative results can still vary.

Will the original audio be retained?+

The prompt requests it, but review the generated result because audio retention must be confirmed from the actual output.

More guides