Performance remake · 07

Tiny Desk AI Video Generator

Upload a lead-singer photo, then add any of the four supporting roles you want to replace. Empty roles use the default cast. Generate a 15-second vertical desk-concert video with sound.

Generated sample

Play the sample to preview the generated performance. Turn on sound to hear it.

AI sample · 480p · 15s · 9:16 · With sound

Generated example using all five role photos. Shot timing and backgrounds differ from the reference.

Photos

Required

Upload the lead singer photo. Add any supporting performers you want to replace; empty roles use the default cast.

Add a clear lead-singer photo. You can also replace the keyboardist, left backing singer, right backing singer and saxophonist in their named slots. Leave any supporting slot empty to use its default portrait. Left and right follow the viewer’s perspective.

Photos stay on this device until you hit Generate. They are resized and stripped of location data in your browser.

Format

Format

Same credits for either format.

Quality

Quality

Refunded automatically if generation fails.

The video will be AI-generated. Please label it as AI when you share it.

0 / 5 photos

20 credits

What is the Tiny Desk AI video trend?

This Tiny Desk AI remake turns your photos into a desk-concert performance. The lead singer needs a photo; keyboardist, left and right backing singers and saxophonist are optional. Each empty supporting role uses its default portrait. The adapted reference guides the performance and generated audio. Shot timing, faces, gestures and instruments can vary, and the original recording is not guaranteed.

Selected sample case

This example was approved for public display. The video appears above.

Input roles
Five photos: keyboard, left and right backing singers, lead singer and sax player.
Generated
2026-11-03
Selected for display
2026-11-03
Sample output
480p · 9:16 · about 15 seconds · generated sound
Current template cost
20 credits at the listed resolution
Tested version
Seedance 2.0 Mini · v4
Editing
Web compression; no cropping, retiming, splicing or face editing.
Observed limits
An extra saxophone shot shortens the lead singer’s part, so the four-shot sequence is not an exact match. We have not verified the sound by listening and cannot guarantee the original recording.

This is one selected example, not a success rate. Total attempts are unknown.

The performance, in four moments.

The adapted reference moves from the band to the lead singer, the saxophonist and back to the band. These are reference timings; generated shots and gestures may differ.

  1. 01

    0–4.29 s · The band

    Keyboardist in front, with two backing singers behind. Each uses its uploaded role photo or its default portrait.

  2. 02

    4.29–9.50 s · The lead

    Your required lead-singer photo guides the solo performance, including singing, laughter and head movements.

  3. 03

    9.50–12.33 s · The sax

    Your optional saxophonist photo, or the default player, takes the solo. Playing posture and instrument interaction use the reference as a guide.

  4. 04

    12.33–15.00 s · The band returns

    The reference returns to the opening keyboardist and backing singers, keeping the same uploaded or default identities.

The original 18.25-second line-art source is retained. The generation reference keeps every shot while fitting 15 seconds. The published sample adds a saxophone angle and shortens the lead segment; gestures, faces and instruments can vary in each result.

Tiny Desk AI questions

The details before you bring your photos.

Can I generate a Tiny Desk AI video now?

Yes. Upload a lead-singer photo and optionally replace any of the four supporting performers. Empty supporting roles use default portraits. The template generates a vertical 15-second video at 480p with model-generated sound. Review the current credit charge before submitting.

Is Tiny Desk AI an official NPR product?

No. RapVideoAI is an independent AI video product. NPR's Tiny Desk Concerts and Charlie Wilson's original performance are the cultural reference, not partners or sponsors. The linked NPR article documents the original concert; it does not endorse this template.

How do I make a Tiny Desk AI video with my photos?

Upload 1–5 photos. The lead singer is required; the keyboardist, left backing singer, right backing singer and saxophonist are optional. Add photos to the roles you want to replace and leave the others empty to use the default cast. Each photo stays assigned to its named role.

Who plays the saxophone and the supporting band?

The keyboardist, two backing singers and saxophonist each have an optional photo slot. Upload a photo to replace that role, or leave the slot empty to use its default portrait. The lead singer always needs your photo. Left and right follow the viewer’s perspective.

Will it include Outstanding or clone my voice?

The adapted reference audio guides performance timing, and the model generates the output sound. This does not guarantee an exact copy of Outstanding or the original recording. Photos do not provide a cloned voice; audio quality and mouth alignment need review.

Will I need to write a prompt?

No. The template supplies the named roles and performance reference. Upload the lead singer and any supporting performers you want to replace; empty supporting roles use defaults. Make your own remains a separate creative option.

Is it free, and how many credits will it cost?

Watching the sample is free. Generation uses credits; the current charge is shown on the create form before you submit. Credit packs are one-time purchases without a subscription.

How long will it take, and what happens if it fails?

Render times and results vary with provider demand. You can return to the result page later. Review identities, expressions, instruments and sound after generation. Technical generation failures automatically refund the credits used.

Whose photos can I use, and can I share the result?

Use photos you own or have permission to use. Do not present an AI performance as authentic or imply endorsement. When adding a soundtrack, follow the sharing platform’s permitted use.