Hotel Lobby AI - Generate two-person orange booth duet videos from two separate photos

Core Functionality

Hotel Lobby AI implements the popular two-person one-mic duet trend without requiring users to edit multiple tools manually. It uses GPT Image 2 to composite the two uploaded portraits into the signature matte-orange recording booth scene, then applies DreamActor motion transfer to add pre-defined performance motion and matching audio to both subjects. The final output is a 720×1280 vertical MP4 file with synchronized audio.

Supported Workflow

The full generation process follows three simple steps:

1. Upload two clear, portrait-oriented photos of authorized adults. Each file can be in JPG, PNG, or WebP format with a maximum size of 10MB. Assign one person to the left side of the microphone and the other to the right side.

2. Add an optional short styling note to define the sequence: who starts first, what gestures to use, and how the performance handoff works between the two people.

3. Run generation. The tool processes the scene in under a minute, and the full video is ready in 3 to 5 minutes. Review the output to confirm faces are recognizable, positions are correct, and no visual artifacts appear before sharing.

Use Cases

The tool is optimized for three common duet scenarios:

• Friend duets: Build call-and-response clips where two friends take turns performing next to the shared microphone

• Sibling back-and-forth: Create playful exchange scenes between family members from separately taken photos

• Couple synchronized clips: Make quiet, shared performance moments for social media content

Tips for Better Results

These steps reduce common generation issues such as face drift, warped hands, and swapped positions:

1. Use sharp, well-lit portraits with fully visible faces. Avoid heavy filters, sunglasses, cropped faces, and crowded backgrounds. Match camera angle and lighting across both source photos for more consistent results.

2. Keep your prompt short and focused. Describe one simple exchange: one person leads, the other reacts, then roles swap. Explicitly state who moves and speaks at each part of the scene, and request a locked static camera.

3. Check your first output carefully. Watch the clip without sound to confirm the handoff action still makes visual sense. If the result has issues, simplify your prompt and run generation again.

Pricing

All preview and setup steps are free. Video generation uses credits that you can purchase via one-time packs or monthly subscriptions:

• Basic: $20 per month, 2000 credits (around 1 duet video), credits refill monthly

• Pro: $50 per month, 5500 credits (around 4 duet videos), credits refill monthly

• Studio: $100 per month, 11500 credits (around 9 duet videos), credits refill monthly

All paid plans output 720×1280 vertical MP4 files with audio. All subscriptions can be canceled at any time.

Privacy Notes

Your photos are resized locally in your browser before being sent to the AI provider fal.ai for generation. Photos are never uploaded before you press the generate button. You must hold full permission to use any portrait you upload, and you are required to label your output as AI-generated when sharing it publicly.