How to Make a Tiny Desk AI Video from Your Photo

You can make a Tiny Desk AI video from a portrait using the homepage studio. The current workflow uses a built-in desk-concert performance template, so you do not need to upload a source video or write a performance prompt. Start with one clear photo of yourself or a person who agreed to appear.

What do I need before I start?

Prepare a JPG, PNG or WebP portrait no larger than 8 MiB. Use a sharp, evenly lit face with the head and shoulders visible. Sign in and verify your email before uploading for generation. A signed-in account also needs enough video credits.

The current template produces a 15-second video and uses one video credit per task. These are the settings checked on October 30, 2026; check the studio again before submitting. The interface has five role positions, but the currently verified input limit is one person. A five-position layout does not demonstrate five-person output quality.

Step 1: Choose the portrait for the main role

Start with the lead singer. Avoid faces partly covered by sunglasses, hair or a hand. Crop out other people rather than asking the generator to choose one face in a group shot. See the photo requirements guide for a checklist.

You can select a file from your computer or use the studio's phone transfer option. Check whether the photo field says it is a local preview or an uploaded photo. Selecting a file alone does not submit a video task.

Step 2: Check the built-in template

Review the reference before spending credits. Template mode supplies the performance direction and 15-second duration. The general scene guides contain creative prompts, but those prompts do not change the active template or unlock another model capability.

Reference examples show the intended style. They are not evidence that your photo has already been rendered, and they do not guarantee identical movement, likeness or sound.

Step 3: Submit once and follow the task

Confirm that you have permission to use the photo, check the credit requirement and submit. Open My videos to follow the task. If it is queued or processing, check that record before starting another task. A slow job is not a reason to repeatedly submit the same photo.

If a task fails, check its status and credit result. For help, send the task ID through support; do not put private photo URLs or account credentials in a public post.

Step 4: Review the result before downloading

A successful task has a playback and download option. Check the face in close-ups, hands, microphone placement and transitions. Listen to the actual output before sharing: exact lip synchronisation, a specific voice and the original concert recording are not promised.

For TikTok or Reels, inspect the downloaded file and choose audio available for your use on the destination platform. This tool does not grant a licence to a song or concert recording. Tiny Desk is an independent AI project with no affiliation to NPR or the original performers.

Is there a free personalised preview?

Do not assume one is included. Reading the guides and viewing reference examples is free; personal generation requires credits. New credit purchases are currently unavailable. If you do not already have usable credits, read the free-versus-paid explanation before preparing a paid workflow.

Open the photo-to-video studio · Read the complete workflow

For both soundtrack options, read the music and silent MP4 download guide.