How to make an Outstanding Tiny Desk AI video
Step by step: cast the band seats, add photos, pay once and download an ~18s 9:16 Outstanding Tiny Desk video with sound. No editing app, template or prompt.
Step 1: Get your photos ready
Use one clear photo per person: one adult, chest-up, face visible, in good even light. Skip sunglasses, group shots and heavy filters. Only use photos of yourself or of adults who agreed to be in the video.
Step 2: Cast the band seats
Open the tool on the homepage and add a photo to each seat you want to recast. Fill one seat or all five; a seat without a photo keeps the original performer. The seats are:
- Lead A: the singer with the big laugh
- Keys B: front row, at the red keyboard
- Backing C: back row, left
- Backing D: back row, right
- Sax E: the saxophone solo
Step 3: Check the credits and render
Every video costs 10 credits, however many seats you fill, and the total is shown before you generate. Sign in, get a one-time credit pack if you need one, tick the consent box and start. Only the shots that contain your people are regenerated; the others keep the original footage. If a render fails, its credits come back automatically.
Step 4: Download and post
You get a vertical 9:16 MP4, about 18 seconds, with sound and no watermark. Post it to TikTok, Reels or Shorts and label it as AI-made. Finished videos stay in My videos in your account.
Do you need a CapCut template or a prompt?
No. Many tutorials point to a CapCut template and manual face swaps, one face at a time. Here the cuts, the timing and the song are built in: you only add photos.
Tips for a better result
AI output varies, and the photo matters most:
- Face not covered by hair, hands, hats or a microphone
- A natural expression works better than a squint or a wide-open mouth
- Recent photos look more like the person today
- For group videos, similar framing across all photos keeps the band consistent