How to Make a Tiny Desk AI Video

1. Choose the roles you want to replace

The appeal of the Outstanding meme is recognizing someone unexpected in the lead. Choose a single, sharp portrait with the face looking toward the camera. Even light and visible head and shoulders give you a clearer subject to work with. Crop a group photo to one person rather than expecting everyone in it to become a different band member.

Use your own photo or one of an adult who agreed to appear. The current uploader accepts JPG, PNG, WebP up to 10 MB. A heavily filtered face, sunglasses or an extreme camera angle can make it harder to recognize the person once the reference motion is applied.

2. Check the shared template and result

A tiny desk ai tutorial may use several tools or hand-built prompts. Here, upload your own photo for selected fixed roles and keep the rest of the template: keyboardist, lead singer, backing vocalist on the viewer’s left, backing vocalist on the viewer’s right and saxophonist. Each photo must show one adult.

Each task produces one video: 15 seconds (15 credits). The five before-and-after comparisons use frames from one shared demonstration video, not five independent outputs. Watch the reference and listen to the demonstration before choosing this template; they do not predict your own result.

3. Sign in and confirm the credit cost

Personal video generation is paid. Signing in or preparing role photos does not unlock a free render. Buy a credit pack or an available plan only after checking the template and cost.

One-time packs are 20 credits for $9.90, 50 credits for $19.90, 150 credits for $49.90. Check both the purchase price and the cost of one video; a pack can leave a remaining credit balance after a take.

The create button shows the required credits. A completed take uses those credits, including when you decide you would prefer another photo. A confirmed failed task returns its credits automatically. A queued task is still in progress; inspect its status before submitting a duplicate.

4. Review, download and share

Return to My videos to follow the task. Watch the whole clip at normal speed and pause on turns or close-ups. Look for a recognizable face, natural-looking motion and distracting changes to clothing or background. Listen to the audio too: the tool does not promise the original Outstanding recording or exact lip sync.

Download the finished MP4 when you are happy with the take. Portraits are deleted automatically 7 days after upload; completed videos are kept for 30 days after they finish. Keep your own copy if you want it beyond that period. When sharing, identify the clip as AI-generated so the joke is understood as an imagined concert.

Following YouTube or Reddit tutorials?

This workflow has five fixed casting controls. Replace or retry any individual slot before submitting. Selecting zero to five replacements creates one 15-second task and uses 15 credits in total; it does not create five separate paid takes.

When comparing a YouTube tutorial or a Reddit discussion, check its actual tool, reference assets and number of character inputs. A multi-person example may involve several generations and editing steps. Compare that process with the five-role, single-video workflow here rather than assuming every tool offers the same controls.