Upload one portrait for Solo or two for Duo, create a short orange-booth performance and review your MP4 here. Start with MiniMax H3 at four seconds and 768p; add your chosen soundtrack after downloading.
Photo 1 = left; photo 2 = right. Use a clear portrait for each.
Two friends singing and rapping at one suspended microphone from the opening moment. Start with two separate clear portraits. Review the generated audio and add your chosen platform sound after downloading.
Compare a real two-portrait result generated here with three original VidLoop scene animations. Inputs and settings appear below each video; sound starts muted.
Generated here from two separate original portrait references. Each performer keeps their own outfit and position, with natural mouth movements and small gestures. This is one reviewed result; outputs vary.
MiniMax H3 · 768 × 1344
Two young adult friends in green and blue share light knee bounces, smiles and low hand gestures. Generated from one complete original scene image.
MiniMax H3 · 768 × 1344
An original golden retriever and black cat sway together, then the cat raises a paw. Generated from one complete original illustrated scene image.
MiniMax H3 · 768 × 1344
An original robot and pink plush alien share a small dance and friendly waves. Generated from one complete original character scene image.
MiniMax H3 · 768 × 1344
The workbench above handles the entire draft: choose a preset, add the required photos, check settings and read the credit quote. Duo assigns the first portrait to the left and the second to the right. Solo uses one portrait. Your image is an identity reference rather than a still-photo opening frame. The prompt asks for the orange scene and natural performance to begin immediately.
Try this setup
When the task finishes, play the result in the workbench. Check the opening frame, each performer’s eyes and mouth, the microphone and hands, and the final frame. A returned video is not automatically a good likeness. The reviewed example below is an actual two-portrait output generated here; the other labeled samples begin with a complete illustrated scene and demonstrate a different input method.
Try this setup
The starter uses four seconds at 768p with vertical 9:16 requested. Our displayed H3 examples measure 768 × 1344 and run about 4.46 seconds, so allow for small provider differences. A longer duration uses more credits; the button and quote show the current cost. Keep small movements and a fixed camera first, then increase the duration only when the result meets your needs.
Try this setupKeep the same account or guest browser identity and return to the workbench’s Recent works. A task that is still processing should be checked through its existing status control. Avoid resubmitting because the first task is slow: uncertain provider tasks keep their reservation until their outcome can be reconciled. Signed-in history belongs to your account and private media cannot be downloaded by another user.
Use the result’s download control after playback. It requests the protected video file rather than sending you to a provider URL. Save the MP4 and open it in a compatible editor or player to confirm the file plays. The visual prompt requests silence but generated audio can remain, so listen before posting. Add licensed music available to your social account and check that your intended use is permitted.
Replace a rejected file with a readable JPG, PNG or WebP under 10 MB. If human verification fails, use the visible retry control and keep the photos in place. A confirmed failed generation returns its reserved credits; a creative redo of a completed result has a new quote. If a task stays unresolved, keep its ID and contact support instead of submitting the same request repeatedly.
Suggested inputs and review checks for your next attempt.
Duo · 4 seconds · 768p · vertical
Check: Both people remain recognizable, with no still-photo intro
Choose the matching scene preset with one image
Check: Image-to-video keeps the intended composition
Replace the file or check the existing task status
Check: No duplicate submission while the first result is unknown
Use one clear portrait for Solo or two separate portraits for Duo. Select a template to fill the prompt and keep each reference in its own upload slot.
Start at 4 seconds and 768p with MiniMax H3. Check the credit quote, then generate. The workbench shows progress and keeps the result available for review.
Play the MP4 and check faces, hands and timing. Download the result, then add a licensed platform sound in your preferred social editor.
Hotel Lobby AI plans
Choose a subscription for regular creation, or a one-time credit pack below. One studio. Three monthly allowances. Start with your free credits, then choose the plan that fits your creative rhythm.
Monthly plans renew automatically. Cancel future renewals from your account.
Your monthly creative allowance
180 MiniMax H3 credits
Your monthly creative allowance
360 MiniMax H3 credits
Your monthly creative allowance
800 MiniMax H3 credits
Video estimates use 5 seconds at 768p (5 credits). Longer clips, 2K output and new variations use more credits. Every plan uses the same MiniMax H3 studio.
Checkout with PayPal
Review your total in USD before you approve payment.
Clear credit costs
See the exact quote before generating. Confirmed failures return eligible credits.
Help from a real person
Credits are added after payment confirmation. Refund policy · Terms · Privacy
1 credit/second at 768p; 2 at 2K. Confirmed failures return eligible reserved credits. Compare full plan details.
A single payment. A balance you can use at your own pace. Valid for 365 days, with no automatic renewal.
Try your first ideas.
80 H3 credits
US$2.99
One-time payment
More room to experiment.
300 H3 credits
US$8.99
One-time payment
Space for your next collection.
700 H3 credits
US$18.99
One-time payment
Billed in USD; credits arrive after payment confirmation. Expiring H3 credits are used first. Failed tasks return eligible credits to their original balance.
Keep creating with tools that fit your source material.
Made something with Hotel Lobby AI? Tell other creators how it went and what you would change.
Be the first to share your experience. There are no published creator reviews for this page yet.
We review feedback before publishing and use only the public name you choose. For help with a task or billing, contact support.
Photos, credits, audio and your finished clip.
Wait for the task to finish, play the preview, then use its download control to save the MP4. Return with the same account or guest browser identity to access private results.
Provider queue times vary. Use Check video status for the existing task or return to Recent works. Do not start the same request again while its outcome and reserved credits are unresolved.
Those are the requested settings. Provider outputs can vary slightly: the reviewed four-second H3 examples here are about 4.46 seconds at 768 × 1344. Check the final file before editing.
Solo and Duo use identity references rather than fixing your portrait as the opening frame. The prompt asks for performance from the first frame. Check the opening and mouth movements; generated motion varies and is not synchronized to a supplied song.
Use one sharp, well-lit portrait per person with the face unobstructed. Similar lighting helps a duo; visible shoulders and hands provide useful context. Upload JPG, PNG or WebP, up to 10 MB per image.
These presets generate a new orange-studio performance from your photos. Faces, gestures and timing can vary. They do not transfer the original choreography, reproduce an exact frame-for-frame performance or guarantee synchronized lips.
The prompt requests a silent performance; generated audio may still be present. After downloading your clip, choose the licensed sound available to your account inside TikTok, Instagram or YouTube. Music availability and permitted uses depend on the platform, region and account.
A vertical MP4 requested at 9:16; our H3 examples measure 768 × 1344. The starter uses four seconds at 768p. Longer clips and higher resolutions need more credits. Review faces, hands and audio before posting. Rendering time varies with the provider queue.
Bring a clear portrait, choose your orange booth direction and make a short first draft. Your finished clip is ready to review and download in the workbench.
Make your videoUse photos you have permission to animate. Review faces and hands before sharing. Hotel Lobby AI is independent of COLORS, the artists and the social platforms.
Photo privacy · Get help · Updated October 6, 2026
