How to Make a Tiny Desk AI Video From Photos

A Tiny Desk AI video starts with a fixed performance and a small set of clear choices. Instead of writing a long prompt, you choose the people or pets to replace, upload one authorized photo for each selected position, and review the request before creating the video. This makes the workflow useful for a group-chat joke, a pet meme, or a private creative test.

What you need before you start

Prepare one clear photo for every performer position you want to replace. The current Outstanding template exposes five positions: Lead, Sax, Keys, Back left, and Back right. You do not have to fill every position. A one-person test is valid, and the remaining performers stay in the reference.

Use photos that you own or have permission to transform. A public image is not automatically licensed for an AI edit. Tiny Desk AI accepts adults or pets only; do not upload minors, public figures, celebrity likenesses, or protected brand assets.

Step 1: Choose the performer position

Open the Tiny Desk AI creator and look at the full reference image. Tap the exact position you want to change. Choosing the position first matters because each uploaded file is mapped to one role. If you want to replace only the lead, select Lead. If you want a complete group meme, select all five positions and prepare five separate references.

The position labels are part of the control, not just decoration. They help you check whether the person or pet is being assigned to the correct part of the frame before a credit is used.

Step 2: Upload one clear photo

Use a readable, front-facing photo with visible facial features or a clear pet face. Avoid a tiny subject in a crowded group image, heavy blur, extreme filters, and aggressive crops. The Tiny Desk AI photo guide has a practical checklist for lighting, crop, and file selection.

After upload, confirm that the selected position visibly changes from an example state to your photo. Check the counter, filename or thumbnail, and the remove control. If the wrong image was selected, replace it before moving on. A clean mapping is more useful than uploading several uncertain references.

Step 3: Review the fixed-template promise

The current workflow is designed around one fixed reference. The room, desk, camera cuts, instruments, timing, and selected output settings are controlled by the template. Your main creative variables are which positions change and which authorized photos are used. This is different from an open text-to-video prompt, where the whole scene may be reinterpreted.

Read the consent statement before submitting. The output is an AI-generated fan-made creative work, not an official performance. If you publish it, follow the disclosure, likeness, music, reference-footage, and platform rules that apply to your use.

Step 4: Create and review the result

The current starter option is shown as one one-time video credit for a 1–5 position generation. Confirm the price in the live checkout before purchase. After submission, wait for the original task to resolve before trying again. If a provider failure occurs, check the task status and the configured credit-refund path rather than creating duplicate requests.

The best first test is usually one position with one strong photo. Once the identity and framing look useful, try a two-person or full-cast version. This keeps the first decision small while preserving the fun of a complete group result.

Ready to test the workflow? Open the creator, or read the pricing and credit guide first.