How to Make the Outstanding Tiny Desk Set With Your Own Photos

Under every viral Outstanding Tiny Desk set, the same questions pile up: "What are people using to make these," "How do you do this??" and our favorite, "how do I create this with our team from work?" This guide answers them: what the famous versions were made with, what it takes to do it by hand, and how to make one with your friends in a few minutes.

What AI made the viral Tiny Desk videos?

The video that started the trend was made with Higgsfield. Its creator, Jake Whipple, told Sport & State he cropped 30 seconds of the NPR clip to vertical, uploaded photos of five coaches and let Higgsfield swap them into the band. When a commenter under his TikTok guessed a different app, he replied that he used Higgsfield and that the other one was "not as good."

Other names that come up in the comments:

Viggle, behind the older "washed" sports edits. One commenter noted "usually people use viggle ai but this one looks significantly better."
Kling, which shows up in a wave of "how to make" tutorial posts.
Summrs, an app whose own account posted a New Orleans Saints version.
CapCut templates, tagged on many reposts. Several commenters said the template didn't work for them; a template reuses someone else's edit rather than casting your own faces.

All of them share one idea, video-to-video character replacement: the motion, camera and timing come from the original clip, and the model redraws each person from a photo.

The do-it-yourself route

If you want to build it by hand the way the first one was made, this is the job:

Get the clip. Find the end of Charlie Wilson's performance in NPR's Tiny Desk video (around 23:40) and cut the seconds you want.
Make it vertical. The original is 16:9. Crop every shot to 9:16 so the faces stay in frame, which is harder than it sounds when the band shot has three people spread across the desk.
Prepare a photo per person. One clear face each, matched to the right musician.
Run the swap and wait. Whipple estimated about a minute of rendering per second of video. Expect a few tries before every face lands.
Put the sound back. Some tools replace or drop the audio, so you may need to lay the original track back under the video in an editor.

Watch out for one trap: editing a video that was already edited. Under one NFL version, the top replies were "I still see the college coaches," because faces from the earlier edit bled through. Always start from the original footage.

The fast route: cast it in the studio

We built Outstanding Tiny Desk AI so none of that is your problem. The clip is already cut, cropped and timed, and the live recording stays on the video. You only bring the photos.

Five made-up coworkers, one photo each. The bearded lead gets the glance and the laugh, and the guy in green flannel gets the sax.

Pick People or Pets. Pets works the same way, with your dog or cat at the desk.
Cast the parts. Upload a photo for each role you want to fill. Only the lead is required; any empty part is played by a made-up stranger. Have one group photo instead? Use Cast from a group photo and tap each face.
Choose Standard or HD. People come out at 720p or 1080p, pets at 480p or 720p. It is 10 credits for Standard and 20 for HD, for one person or five.
Wait five to ten minutes. You get a 15-second, 9:16 MP4 with no watermark, saved in My videos and ready to post from your phone.

Packs start at $9.90 for 10 credits with no subscription; see credit pack pricing. If a render fails, its credits come back on their own.

Who to cast where

The comments make it clear which parts carry the joke, so cast with that in mind.

The lead is the star. The lead plays both close-ups, the glance at 0:00 and the laugh from 0:07 to 0:13, about seven of the 15 seconds. The most-liked comment on almost every version is some form of "[lead] singing his heart out." Give it to the person being roasted, the one who lost the bet or the birthday person.

The sax is the punchline. The solo is only the last two seconds, and it gets the biggest reaction: "I lose it everytime at whoever gets this part." Save it for the person nobody expects, like the quiet coworker, your grandpa or the boss. Give them a photo turned slightly to the side if you have one, because the sax close-up is the hardest face to match.

Keys and the backup singers set the scene. They share the band shot from 0:01 to 0:07. Keys also peeks into the corner of the sax shot, so it is a good spot for whoever should be watching the reveal.

Let the outfits do the work. Viewers loved that the coaches "were still wearing their headsets," and they noticed when a coach's signature visor was missing. In your video the clothes come from the photo, so take it in the jersey, the work lanyard, the headset or the hat that everyone knows you for. Belt-up photos bring the outfit along; face-only selfies do not.

"Can someone make one for my team?"

It is the most common request in the comments: Jets, Steelers, Raptors, a WWE version, a Barstool version. On our site, every photo has to be you or an adult who agreed to be in it: no coaches, players, celebrities or other public figures, and no one under 18.

You can still make it about your team. Shoot the watch-party version: you and your friends in team gear, cast as the band that has to sing through another loss. The joke still works, because everyone in the group chat knows exactly who is on the sax.

Before you post

Label it as AI. The first one said "This is an AI parody!" right in the caption, and it still went viral.
Expect the sound to get muted on some apps. The song belongs to its owners. If TikTok or Instagram mutes it, add "Outstanding" back from the app's own sound library.
Keep it personal. It is a fan parody for you and your friends, not an ad.

Want to know where the clip comes from and why it works? Read the Outstanding Tiny Desk meme, explained. If a face in your video comes out wrong, these fixes for faces, age and blocked photos walk through what to change.