Where does a Mi Chico Dance result most need a second look?

We have not measured failure rates. The moments that ask the most of a generation are the opening huddle, where five faces and many hands crowd the frame, the blurred phone hand-off at about 4.5 seconds, and the routine, where dancers swap places as the formation rotates.

Why are some faces upside down at the start?

That is the shot, not a fault. The phone lies on the ground looking up, so the dancers at the top of the frame lean over it and appear upside down.

What do I change first if one friend does not look like themselves?

That friend's photo. Use a sharper, front-facing photo with the face clearly lit and the outfit visible before changing anything else.

Should everyone send the video before it is posted?

Yes. Five people appear in it, so share it with the group first and let anyone who is unhappy with how they came out say so.

Check Your Mi Chico Dance AI Video Result

Your Mi Chico Dance video is back. Five people are in it, so there is more to check than in a solo trend. Watch it once at normal speed, then step through the beats below. This is the review half of the pair; the how-to guide covers choosing the five photos in the first place.

Make or remake your crew video

Open Mi Chico Dance, add five photos, review the quote and generate the 17-second rooftop crew video.

Create this trend

Before you start: what "right" looks like

The template follows a 17-second take by @jyt with no cuts. Your result should keep that structure: huddle, selfie walk-back, a blurred hand-off, the routine on a rooftop car park at sunset, hands on heads, a whip-pan to the sky. The preview on the trend page is the credited source, not an Oakgen output, and we have not reviewed a generated example yet, so the checks below are what to look for rather than a description of typical results. The shot-by-shot breakdown has the timings.

Beat 1: the huddle (0 to 1.5 seconds)

Credited source frame at about 0.4 seconds: five faces lean over a phone on the ground against a blue sky

The opening huddle in the credited @jyt source video, not an Oakgen output.

This is the only moment all five faces are close to the lens. Pause on it and check:

Five people, five different faces. No one should appear twice and no face should be a blend of two friends.
The right person in the right place. The leader is bottom centre, the blue beanie lower left, the black cap right, the sunglasses and graphic tee at the top. The top two are upside down because they lean over the phone; that is the shot.
Hands. Waves, peace signs and reaching hands crowd this frame. Count fingers on the hands nearest the lens.

Beat 2: the selfie walk-back (1.5 to 4.4 seconds)

The leader picks up the phone and films themselves walking backwards with the crew behind. This is the leader's close-up, so check their face, hair and the outfit from their photo. The phone itself should stay in their hand, not double or vanish.

Beat 3: the hand-off (about 4.5 seconds)

The phone swings down in a blur as the leader steps into the formation. The source is longer than one AI video section, so the template joins two sections at this blurred moment. Watch it at normal speed and then frame by frame: look for a dancer who changes clothes, swaps face with a neighbour or jumps position across the blur. A short blur is expected; a person changing identity is not.

Beat 4: the routine (4.6 to 15.5 seconds)

Credited source frame at about 10 seconds: the crew dances on a rooftop car park at sunset with light poles behind, a dancer in a black cap and one in a blue beanie in front

Mid-routine in the credited @jyt source video, not an Oakgen output.

The formation rotates, so different dancers come to the front while the camera moves around them. Pick one friend at a time and follow them through the routine:

Do they keep the same face, hair and outfit every time they come forward?
Does anyone multiply, so six dancers appear, or disappear behind the others for good?
Do arms and hands hold together through the chest pumps, crossed-arm sway, jabs and roll?

The light changes from sunset glow to lit light poles during this section. That is in the source, not a fault.

Beat 5: the finish (15.5 seconds to the end)

Credited source frame at about 15.5 seconds: the crew stands with both hands on their heads on the rooftop at dusk under a light pole

The hands-on-head finish in the credited @jyt source video, not an Oakgen output.

All five put both hands on their heads, then the camera whips up to the sky and a light pole. The hands-on-head frame is the best thumbnail candidate, so check it closely: five dancers, ten hands, each on the right head.

Sound

Your video keeps the source's audio: the TikTok sound "Official Mi Chico", credited to Jason Derulo, from the Jason Derulo and Melody version of DJ Goja's "Mi Chico". Check that the routine still lands on the beat. If you post to a platform that lets you pick a sound, keep the original rather than replacing it with a different edit of the song.

Keep it or redo it

What you see	Most likely fix
One friend does not look like themselves	Replace that friend's photo with a sharper, front-facing one
Two friends look alike	Use photos with more visible differences in hair or clothing
An outfit is not what the person wanted	The photo decides the outfit; upload one in the right clothes
Faces swap across the hand-off	Regenerate; check whether two photos are very similar

Before posting, send the video to all five people. Say in the caption that it is AI-made, and if you show the source alongside, credit @jyt. Readers who arrived from the cat and dog versions can find the background on the crew vs animal versions page.