A small tool for one fifteen second joke

Tiny Desk AI does one thing: it puts your people into the desk concert shot that everyone has been reposting since the middle of October 2026. You send one photo or five, pick a mode, pay once, and fifteen seconds of video comes back with the sound already in it.

We are not a video editor and we are not trying to become one. There is no timeline, no prompt box, no library of filters. The shot is fixed, which is the whole point: nothing about the clip has to be invented, so nothing about the render can go dramatically wrong except the photos you send.

What this is not

It is not connected to NPR, NPR Music, the Tiny Desk Concerts series, or anyone who has ever played one. No endorsement, no sponsorship, no partnership. The name describes the format, nothing else, and the videos are made by a model rather than by a camera.

It is also not a way to make a real person appear to have said something. If you put a public figure in a video and post it as real, that is on you, and it is the fastest way to lose an account.

How the money works

One video is $4.99, three are $9.99, five are $14.99. Every order is a one-time payment and nothing renews on its own. There is no free trial because every render is paid GPU time, and we would rather keep the first video cheap than hand out renders that mostly go unused.

If a render comes back wrong, reply to the receipt within seven days. We refund it, and the note you send about what went wrong is also what makes the next attempt better.

What happens to your photos

Photos go into private storage, are used only to render your video, and are deleted within 24 hours of the render finishing. They are never shown publicly and never used to train anything. The video itself stays downloadable for seven days, so save it the same day it arrives.

If you would rather not upload a photo at all, that is fair, and there is nothing else this tool can do for you. Everything here runs on the face you send.