How to make the Outstanding Tiny Desk AI video
Three ways to put your friends into the viral "Outstanding" Tiny Desk concert – with or without CapCut – and what to do when the result looks wrong.
What you are making
The trend is a 15–19 second vertical clip: a soul band behind a cluttered office desk sings "Outstanding" (The Gap Band, from Charlie Wilson's Tiny Desk concert), the lead singer laughs with his head thrown back, and a sax player takes a solo. Every version keeps the set, the order of shots and the sound, with new people in the band. See the shot-by-shot trend guide for the details.
3 ways to make it
1. A generator made for this trend (no editing)
Upload a photo of the lead singer, plus up to four bandmates if you want, and the AI recreates the scene and the order of shots with your people, with the trend's sound already on it. It is generated, not a frame-perfect copy: faces follow your photos closely but not perfectly, and moves differ a little from the original. This is what OutstandingTinyDeskAI.com does: 15 seconds, 9:16, from $9.90 one-time for a Mini 480p video, no subscription and no app.
2. A template app
Some AI video apps add the trend as a template. For example, Summrs has a "Charlie Wilson Tiny Desk" template that swaps five faces into the original Tiny Desk performance (all five photos required; about 19 seconds, 480p or 720p, with the performance audio). It works on credits: its site data lists the template at 430 credits per video, while sign-up gives 50 free credits and the $15/month plan includes 300.
3. A face-swap tool on the original clip
General face-swap tools such as Remaker AI can swap several faces in a video you upload. You have to find and trim the source clip yourself, assign each face, and add the sound. It gives you the most control, but it is the most work.
Side by side
| OutstandingTinyDeskAI | Template app (Summrs) | Face-swap tool (e.g. Remaker AI) | |
|---|---|---|---|
| You upload | 1–5 photos | Photos | The source clip + face photos |
| Made for this trend | Yes, only this trend | One template among many | No, general tool |
| People in the band | 1–5 photos: lead, 2 back-up singers, keys, sax (empty roles are AI musicians) | Exactly 5 photos: lead, 3 back-up dancers, sax | Every face you assign |
| Video | New AI-generated scene, 15 s, 9:16, 480p–1080p | Faces swapped into the original footage, about 19 s, 9:16, 480p or 720p | Faces swapped into your clip |
| Outfits | Their own clothes or one outfit you type | From the photos | From the clip |
| Trend sound | Included (first 15 s of the original recording) | Original performance audio | Whatever is in your clip |
| Cost of one video | Mini 480p: $9.90 (Solo pack). 2.0 720p or 2.5 480p: $29.90 (Band). 2.0 1080p or 2.5 720p/1080p: $99 (Tour, 160 credits, leftovers stay) | 430 credits per video (per its site data). 50 free sign-up credits advertised; $15/month for 300 credits or $5 for 50 | Credit packs from $5.99 for 200 credits; per-video cost depends on length |
| Subscription | No | Optional | No |
Other tools' details are from their public websites as of October 24, 2026 and may change – check their sites before you buy. We have not tested their output. Our credits never expire, so a bigger pack keeps the rest for later videos.
Can I do it in CapCut?
Partly. We could not find an Outstanding Tiny Desk template in CapCut's template library (October 24, 2026). CapCut is an editor: it is great for adding captions, trimming or swapping the sound, but it does not put your friends into the band. The easy route is to make the clip with an AI tool, then open it in CapCut or TikTok to add text and post.
When the video looks wrong
- The face doesn't look like them: use a sharp, front-facing photo in even light, one person per photo, no sunglasses or beauty filters. Mini 480p is for testing; Seedance 2.0 or 2.5 keep faces closer to the photo.
- The wrong person got the close-up: the lead slot gets the intro and the long laughing close-up – put your star there.
- Clothes look random: type one outfit for the whole band, like "matching red tracksuits" or "Christmas sweaters".
- Odd hands or teeth: a common AI artifact. Try a different, clearer photo or a higher model.
- The video failed: sometimes the AI refuses a photo. A failed video never costs anything – its credits come back to your balance automatically, no need to ask. Try another photo.
- It finished but is clearly broken (a face missing or on the wrong person, a glitched clip): email us within 7 days. We check it and return its credits. Results that work but that you don't like are not refunded, so test with Mini 480p first.
- The sound got muted on TikTok or Instagram: mute the video's audio and add the official "Outstanding" sound from the app's music library.
Before you post
About the sound: it is a short clip of the original "Outstanding" recording, included for personal, non-commercial posts. We don't license the music, so TikTok or Instagram may mute it; for brand or paid posts, swap in the app's own "Outstanding" sound.
Your photos: used only to make your video, sent to our AI provider (Kie.ai, running ByteDance Seedance) for that, never used to train AI, and deleted after 7 days. Details in the privacy policy.
Only use photos of people who agreed to it, never of minors, and label it as AI. Don't use it to put words in real people's mouths or to fake real events.