Two-photo original rap duet / H3 Max

Hotel Lobby AI Generator

Upload two authorized portraits and direct a new two-person rap duet. Choose who performs on the left and right, write one short exchange, and generate the footage and sound together.

This creates an original performance. It is not the official music video with replaced faces, and it does not include the commercial song, recognizable lyrics, artist voices, or copied choreography.

2 photos · 5–15 seconds · 9:16 MP4 with generated sound

Your original two-photo rap duet

9:16 MP4 · 5–15 seconds · with sound

1 Choose an original setup

Original scene briefs

2 Add two adult performers

Use two separate, clear portraits of consenting adults. The photos guide identity; the prompt creates a new scene.

Photos are uploaded to site storage at provider-readable URLs when selected and sent to the video provider only after you submit a generation. Do not upload sensitive images.

Ready to adapt

Image 1 = left adult performer. Image 2 = right adult performer. 740/2000

Reference mode: 8 credits/s at 480P · 12 credits/s at 768P.

Sign in to save videos and manage credits. You can explore the briefs first.

Before you generate

Scene plan10s · 480P · 9:16

One mic, two verses

A restrained two-person performance built around one mic.

Image 1 Left adult performer

Image 2 Right adult performer

Two authorized adult identity references and an original scene brief—not footage from an existing music video.

The intended beat

  1. 01Find the rhythm
  2. 02Trade one line each
  3. 03Land the final look
An original warm-orange recording booth
An original minimal beat, room tone and two short vocal phrases

1Add a photo for each character

One original vertical rap-duet scene with generated sound. It does not insert the photos into an existing music video or include a commercial recording or artist voice.

Read the Hotel Lobby prompt guide

View My Videos and task history →

A usable first render

Plan ten seconds before you write the prompt

A short performance needs sequence, not a long list of visual adjectives. This four-beat plan gives each person a clear turn and leaves the model fewer chances to lose a face.

  1. 0–2s

    Settle into the frame

    Keep both performers visible while the original beat and room tone establish the scene.

  2. 2–5s

    Image 1 takes the first turn

    One short original phrase and one readable gesture are enough to establish the left role.

  3. 5–8s

    Image 2 answers

    Write a one-beat pause before the right performer replies so the voices do not overlap.

  4. 8–10s

    End on one shared reaction

    A nod, smile, or simple fist bump gives the clip a clean edit point without crossing positions.

Upload order matters

Two photos. Two fixed roles.

Use a separate, clear portrait for each consenting adult. Image 1 guides the performer on the left; Image 2 guides the performer on the right. Keep those labels unchanged throughout the prompt.

Choose references the model can read

  • One face per image, similar face size in both photos
  • Front or slight three-quarter view with visible hair
  • Natural detail without sunglasses or strong beauty filters

Keep the first motion modest

  • Fixed medium-wide camera with both faces visible
  • No crossing positions or fast camera orbit
  • One nod, open hand, or simple shared reaction

Know the workflow

A new scene, not a fixed-video face swap

H3 Max uses the two photos as identity references and interprets your written direction. It does not composite those faces onto source footage frame by frame, so the camera, movement, timing, and audio are newly generated and can vary between attempts.

What this makes

  • A new 5–15 second vertical MP4
  • Two photo-guided adult performers
  • Editable action plus generated beat, voices, and ambience
  • 480P drafts or 768P output

What it does not copy

  • The official footage, camera path, or choreography
  • A commercial song, recognizable lyrics, or artist voice
  • Public figures or people who did not consent
  • A guaranteed face composite in every frame

Original two-person setups

Start with a reason for the duet

A birthday line, team win, or family joke gives the exchange a point. Load a setup, then rewrite the short phrases so they belong to the two people in your photos.

Read the prompt guide

Orange booth duet

One mic, two verses

A restrained two-person performance built around one mic.

Load this setup →

Friendship message

Birthday bars

Turn a birthday line into a compact back-and-forth verse.

Load this setup →

Diagnose the first render

Change one variable, not the whole prompt

Review the silent picture first, then the sound. A specific fix is easier to judge than a complete rewrite.

The performers swap sides

Put the left/right mapping in the first sentence, keep the camera fixed, and remove crossing movement.

The faces begin to blend

Use two separate front-facing portraits with similar face size and fewer filters or obstructions.

Hands become distracting

Replace pointing, finger counting, or a complex handshake with a nod or one open-palm gesture.

Both voices speak together

State the order explicitly: Image 1 speaks, pause for one beat, then Image 2 answers.

Price before generation

A 5-second 480P clip starts at $2.99

Choose the length and resolution in the studio to see the exact credit cost and single-video checkout price. Checkout adds the required credits and restores this draft; it does not generate the clip until you return and press Generate.

See all credit options

Hotel Lobby AI generator questions

What is a Hotel Lobby AI generator?

It is the name people use for a two-photo AI workflow that creates a short two-person rap performance. This generator uses the portraits as identity references, then creates new footage, movement, vocal phrases, and sound from your direction.

Does it put my faces into the original music video?

No. This is H3 Max reference-to-video, not a fixed source-video face swap. It creates an original vertical scene and does not provide the official video, recording, lyrics, choreography, camera path, or artist voices.

How many photos do I need?

Upload two separate, clear photos of consenting adults: Image 1 for the left performer and Image 2 for the right performer. A photo guides identity and appearance; it is not promised as the opening frame.

Is the Hotel Lobby AI generator free?

There is no universal free-generation promise. The page shows the exact credit cost and one-video checkout price before generation. A 5-second 480P clip starts at $2.99, and no subscription is required.

Does the generated clip include the Hotel Lobby song?

No. The generator asks for a new beat, original phrases, and room sound. If a social platform licenses a commercial track for your intended use, add it through that platform after reviewing its terms.

Why can faces or hands change during the video?

Reference-to-video guides a new generation rather than compositing fixed faces frame by frame. Clear portraits, fixed left and right positions, a steady camera, and small gestures reduce drift, but they cannot guarantee perfect identity or hands in every frame.

What settings are available?

Choose 5–15 whole seconds at 480P or 768P. The dedicated two-person workflow makes a 9:16 MP4 with generated sound.

What happens after checkout or a failed generation?

Single-video checkout adds the credits for the selected clip and returns you to the saved draft; you still review it and press Generate. A confirmed provider failure returns the reserved credits, which is not an automatic Stripe payment refund.

Independent third-party tool. Not affiliated with, endorsed by, or sponsored by Migos, Quavo, Takeoff, Quality Control Music, COLORS, or the owners of the “Hotel Lobby” recording and performance video. Use only authorized adult photos and content you have the right to publish, and disclose realistic results as AI-generated.