Tiny Desk AI: A small-room performance with your own cast
Build an intimate music performance against packed bookshelves. Choose a lead singer alone, add a friend on saxophone or create the three-person band version with a keyboard player.
The available You option replaces the selected performer in the full preview while retaining its shot order, movements and the concert soundtrack heard in this template’s preview. It does not ask the model to invent a new song or dialogue. Review the result for identity and lip timing; exact frame-by-frame fidelity is not guaranteed.
Upload one clear adult portrait for the lead singer; the other musicians stay as they appear in the preview, with permission from each person. Duo and Band are available on the Create page; add one clear photo for each selected role.
Example
A closer look at Tiny Desk AI
The preview demonstrates the source performance and soundtrack. Personalized results can vary; review identity, movement and lip timing before sharing.
Visual reference
A visual reference, not a Smew customer result. Your generated details will differ.
Choose a version, add the required photos and check the displayed credits before generating.
Choose the available option
Choose You and upload one clear adult portrait for the lead singer; the other musicians stay as they appear in the preview. Duo and Band are available on the Create page; add one clear photo for each selected role.
Review and generate
The current You option costs 126 credits for approximately 10 seconds at 720p in vertical 9:16. The displayed preview performance and its soundtrack guide the result. No prompt or voice recording is needed.
Review the finished result
Follow the task on Dashboard and open the result in Library. Watch every shot, check each face and speaking role, and listen for sound alignment before downloading or sharing. Processing time varies.
Choose a photo that keeps the subject recognizable
Use one consenting adult aged 18 or over per photo, chest-up with a sharp front-facing face. The lead's close-up makes identity especially important. Even lighting and uncovered eyes are more useful than a dramatic concert pose or a cropped screenshot.
Photo one is always the singer. Photo two supplies the sax player in Duo or Band, and photo three supplies the keyboard player in Band. A group picture in one slot does not assign the instruments. You also do not need photographs for the generated supporting musicians.
Show recognizable hair and clothing, but do not rely on fine printed text or a team logo remaining exact. Avoid sunglasses and strong face filters. All participants should agree to the fictional band joke; a public photograph of a coach, player or celebrity is not a substitute for that permission.
Where the format comes from
The source context connects the meme to Charlie Wilson's 2023 desk-concert performance of the Gap Band's Outstanding. A close ensemble, expressive lead and prominent sax moment provide the recognizable concert grammar. The real performance and later AI parodies are different works.
October 2026 sports-themed examples recast a band using familiar figures, then other templates offered photograph-driven versions. The linked posts explain that background without making public figures a supported upload category here or claiming that any one post invented every desk-band parody.
Smew creates a new desk-concert scene from the chosen adult identities and generated bandmates. It is not an official NPR or artist product, and the current sound is generated rather than an attachment of the original song. The source credits help explain the joke, not grant music rights or promise exact instrumental performance.
The real Charlie Wilson concert supplies musical and performance context. Credited example
A sports-themed AI parody adapts the desk-band format. Credited example
Another sports example changes the cast while retaining the concert joke. Credited example
Photo-template tutorials offer a one-image route into the band format.
Further discussion uses several desk-concert search labels; unverified search rankings are omitted.
Copy and paste
What the built-in recipe controls
These instructions are included in the template. Your uploads provide the identities; the selected version controls the scene and movement.
You
1 photo: you, the lead singer (18+), face clearSeedance 2.510 s · 720p · 9:16126 credits
Video prompt
Edit [Video 1]. Replace only the lead singer with the face and hair identity of [Image 1]. Keep the saxophone player and all other band members, their faces and their performances unchanged. Never put the uploaded face on another band member or an instrument.
Preserve the complete source video from its first frame to its last: original frame order, every shot boundary and its exact timestamp, playback speed, camera movement, framing, background, lighting, clothing, body positions, props and non-target people. Replace the target identities in place; do not recreate the scene from a new storyboard.
Preserve the original performance: head movement, gaze, facial expressions, hand gestures and body action. Preserve each person's original mouth and jaw opening and closing, lip shapes and their exact frame timing. Only the person who speaks or sings in the source may speak or sing visually. Do not invent a new performance, dialogue, lyric or choreography.
Do not add, remove, reorder, shorten, extend or retime shots. Keep the output silent; the matching original soundtrack is restored separately after editing. No new music, speech, captions, logos or watermark.
2 photos: photo 1: the lead singer (18+), face clear + photo 2: the sax player (18+), face clearSeedance 2.5 Video Edit10 s · 720p · 9:16126 credits
Video prompt
Edit [Video 1]. Replace only the explicitly mapped source subjects in place:
[Image 1] replaces the lead singer holding the handheld microphone in the foreground, including his close-up, in every shot where that subject appears.
[Image 2] replaces the saxophone player in the source band; track this same musician when the camera reveals the saxophone solo, in every shot where that subject appears.
[Image 1]: Copy the uploaded person's face, facial structure, eyes, skin tone, age, hair and facial hair or absence of facial hair. Remove source facial hair or hairstyles that the upload does not have. Keep identity consistent across every shot. Keep the source outfit unless the selected variant explicitly changes it or requires a body/species change.
[Image 2]: Copy the uploaded person's face, facial structure, eyes, skin tone, age, hair and facial hair or absence of facial hair. Remove source facial hair or hairstyles that the upload does not have. Keep identity consistent across every shot. Keep the source outfit unless the selected variant explicitly changes it or requires a body/species change.
Selected variant: Duo / duo. Apply only the requested cast, wardrobe, setting or effect changes for this variant; preserve the source action and shot timing instead of inventing a new scene sequence.
Replace only the two mapped adults, with separate identities. Keep the keyboard player on the left and all other bandmates unchanged. Preserve the original saxophone, connected mouthpiece, microphone, hands, office bookshelves, camera cuts and each person's original performance and mouth timing. Never put the lead singer's uploaded face on the sax player.
Track each source subject through front, profile, back and occluded views. Keep all non-target subjects unchanged. Never duplicate an uploaded identity, merge subjects, swap their roles or transfer an identity to a prop or background subject.
Preserve the complete source video from its first frame to its last: source duration, frame order, every shot boundary, playback speed, camera movement, framing, lighting and original performance; keep background/props unless the selected variant explicitly changes them. Preserve each subject's original head, gaze, body, hand/paw gestures, expression, mouth and jaw opening/closing and their exact source timing. Do not create a new dance, scene, shot list, dialogue or lyrics. Do not add, remove, reorder, shorten, extend or retime shots or actions.
Produce silent video. The matching unmodified original soundtrack is restored separately after this edit. Do not generate speech, singing or music. Do not add new captions, subtitles, logos or a watermark.
3 photos: photo 1: the lead singer (18+), face clear + photo 2: the sax player (18+), face clear + photo 3: the keyboard player (18+), face clearSeedance 2.5 Video Edit10 s · 720p · 9:16126 credits
Video prompt
Edit [Video 1]. Replace only the explicitly mapped source subjects in place:
[Image 1] replaces the lead singer holding the handheld microphone in the foreground, including his close-up, in every shot where that subject appears.
[Image 2] replaces the saxophone player in the source band, including the saxophone solo, in every shot where that subject appears.
[Image 3] replaces the keyboard player at the red keyboard on the LEFT of the opening wide shot, in every shot where that subject appears.
[Image 1]: Copy the uploaded person's face, facial structure, eyes, skin tone, age, hair and facial hair or absence of facial hair. Remove source facial hair or hairstyles that the upload does not have. Keep identity consistent across every shot. Keep the source outfit unless the selected variant explicitly changes it or requires a body/species change.
[Image 2]: Copy the uploaded person's face, facial structure, eyes, skin tone, age, hair and facial hair or absence of facial hair. Remove source facial hair or hairstyles that the upload does not have. Keep identity consistent across every shot. Keep the source outfit unless the selected variant explicitly changes it or requires a body/species change.
[Image 3]: Copy the uploaded person's face, facial structure, eyes, skin tone, age, hair and facial hair or absence of facial hair. Remove source facial hair or hairstyles that the upload does not have. Keep identity consistent across every shot. Keep the source outfit unless the selected variant explicitly changes it or requires a body/species change.
Selected variant: Band / band. Apply only the requested cast, wardrobe, setting or effect changes for this variant; preserve the source action and shot timing instead of inventing a new scene sequence.
Replace only these three mapped adults, each with their own uploaded identity. Keep all remaining bandmates unchanged. Preserve each musician's instrument and hands, the saxophone mouthpiece, microphone, office bookshelves, all source camera cuts and each original performance and mouth timing. Never blend identities or move a face to the wrong instrument.
Track each source subject through front, profile, back and occluded views. Keep all non-target subjects unchanged. Never duplicate an uploaded identity, merge subjects, swap their roles or transfer an identity to a prop or background subject.
Preserve the complete source video from its first frame to its last: source duration, frame order, every shot boundary, playback speed, camera movement, framing, lighting and original performance; keep background/props unless the selected variant explicitly changes them. Preserve each subject's original head, gaze, body, hand/paw gestures, expression, mouth and jaw opening/closing and their exact source timing. Do not create a new dance, scene, shot list, dialogue or lyrics. Do not add, remove, reorder, shorten, extend or retime shots or actions.
Produce silent video. The matching unmodified original soundtrack is restored separately after this edit. Do not generate speech, singing or music. Do not add new captions, subtitles, logos or a watermark.
You pay in credits. The Generate button shows this price before you spend anything. These are the options on this page.
Option
Model
What you get
Credits
You
Seedance 2.5
10 s · 720p · 9:16
126
Duo
Seedance 2.5 Video Edit
10 s · 720p · 9:16
126
Band
Seedance 2.5 Video Edit
10 s · 720p · 9:16
126
One-time packs and plans are on the pricing page. A failed video gives its credits back.
A short session with a surprise sax moment
The setting is deliberately close: a desk, bookshelves, records and warm room detail crowd the performers into a small concert space. The lead begins at the microphone, leaning into an expressive note before a laugh-like reaction and hand gesture make the performance feel playful rather than solemn.
The sax passage redirects attention. In You, that player is generated; in Duo or Band, the second uploaded adult is intended to become recognizable at that moment. Band also gives the third adult a keyboard role. The final group gesture closes the session without requiring a full-length song.
The current take is 10 seconds at 720p in vertical 9:16. Seedance generates the scene and audio, with a soul or funk-like performance direction rather than an exact attached recording. Faces, hands, instruments and precise musical timing can vary, so review the solo as well as the lead's opening close-up.
One star, a sax-playing friend or three named bandmates
You is the simplest cast: one uploaded adult sings lead with generated musicians around them. Duo makes the sax reveal personal by supplying a second recognizable adult. Band adds the keyboard identity through a third photo. The recipe still aims for five performers, with fictional bandmates filling the other positions.
The available You option replaces the selected performer in the full preview while retaining its shot order, movements and the concert soundtrack heard in this template’s preview. It does not ask the model to invent a new song or dialogue. Review the result for identity and lip timing; exact frame-by-frame fidelity is not guaranteed. Duo and Band are available on the Create page; add one clear photo for each selected role.
No performance recording is required from the participants. The result imagines them playing and singing, so it is not evidence of actual musical ability or a real office concert. Confirm that everyone enjoys the role assignment before sharing a group-chat or workplace joke, and keep the AI disclosure visible.
Outstanding is the reference context, not the promised recording
The Gap Band song Outstanding connects the real concert and many meme posts. That musical recognition helps explain why a short sax passage or a large vocal reaction feels familiar. It does not mean every AI desk-band tool uses the original recording.
The available You option replaces the selected performer in the full preview while retaining its shot order, movements and the concert soundtrack heard in this template’s preview. It does not ask the model to invent a new song or dialogue. Review the result for identity and lip timing; exact frame-by-frame fidelity is not guaranteed. Duo and Band are available on the Create page; add one clear photo for each selected role.
If you later add a licensed social sound in another editor, check the platform's terms and alignment, especially for commercial publishing. Template access is not a song license. The generated room and band should not be described as an official NPR session or an endorsement by the source artist.
Give the fictional band a personal introduction
A caption can name an office team, a friendship group or a collection of people who have had a difficult week, without claiming the scene really happened. Explain who is on sax if that recognition is the payoff. A short original introduction often gives enough context for the 10-second performance.
Review the band before posting: the singer's face at the microphone, the sax player's close-up and the keyboard player's hands can each introduce a different artifact. The five-person composition means some faces are fictional, so do not tag generated bandmates as real people.
Keep subject permission and AI labeling. Add any desired text in a separate publishing editor rather than assuming the recipe contains custom captions. Descriptive desk-concert tags identify the format, but changing social rankings are not a promise of reach. The strongest joke is a recognizable, consenting cast taking a tiny performance space far too seriously.
FAQ
Tiny Desk AI questions
What is the tiny desk AI trend?+
It is a generated desk-band parody with a highly expressive lead and a surprise instrumental moment, inspired by the concert references linked above. Smew casts your uploaded adults in up to three named roles and fills the remaining positions with generated performers. It is not a real concert appearance.
What song is the Tiny Desk meme?+
The personalized You result retains the concert soundtrack heard in this template’s preview. It does not create a new vocal take. The preview identifies the actual sound used; template access does not grant separate music publishing rights.
Can I put my friends in the outstanding tiny desk set?+
Yes. Duo adds photo two as the sax player, and Band adds photo three at the keyboard. Photo one remains the lead singer. Use separate images of consenting adults. There are no extra photo slots for the other generated performers in the five-person composition.
Can I make it with coaches, players or celebrities?+
Use yourself and adults who have agreed to the depiction. Sports-themed source parodies explain the meme's background, but do not establish permission to reuse public figures' faces. Keep the generated interaction clear rather than presenting it as a real performance or endorsement.
Is there an outstanding tiny desk AI template or tutorial?+
No. Select You, Duo or Band, add the ordered photos and confirm the 126-credit choice. The scene recipe is included. Dashboard tracks processing and Library holds the completed file. Generation time varies rather than following a guaranteed tutorial countdown.
Is this an AI Tiny Desk concert generator with the real song?+
The personalized You result retains the concert soundtrack heard in this template’s preview. It does not create a new vocal take. The preview identifies the actual sound used; template access does not grant separate music publishing rights.
Why does my clip have only five people?+
The recipe aims for a five-person ensemble. Your images fill the lead, sax and keyboard roles according to the selected choice; generated characters fill the remaining places. You can supply at most three identities here. The cast size is a scene design, not a count of additional real people you must photograph.
