Fergie Baby AI With Friends: Two People in One Clip

The most common question we get about group clips is simple: can I put me and my friend in the same video? The honest answer is that you can try, and sometimes it works, but it is not what the tool is built for, and you should know why before you spend credits on it.

This article covers the single-person default, the two-person exception and its risks, the more reliable two-clip route, and the consent rules that apply whenever someone else appears in your render.

The honest answer first

Fergie Baby AI is built around ONE person per photo. The reference clip that drives the animation encodes a gait and a camera move for a single subject walking. That is the whole engine.

When you upload a two-person image, the model has to do something it was not designed for: decide which body owns the walk, and then apply one movement cycle to two different poses. Sometimes the result is perfectly usable. Sometimes one person walks while the other slides, freezes, or deforms, or the two bodies merge oddly at the edges. There is no switch that makes this reliable, and we will not pretend there is.

So the practical rule is: one person for a clip you care about, two people if you are experimenting and willing to accept a variable outcome.

Why one person per photo is the reliable path

A single-subject photo gives the model an unambiguous job. It reconstructs one body, applies one walk cycle, and tracks one head. Every part of the pipeline agrees about who is on screen, so the render is consistent and the face stays stable across twelve seconds.

With two people, all of that becomes a set of choices the model makes on your behalf. Which subject is the primary? Where does the camera move relative to? How should the second body respond when the first one takes a step? None of those have a defined answer, so the model picks something, and what it picks can differ between two renders of the same upload.

That variability is the real cost. It is not that two-person clips never work; it is that you cannot count on one, and if you are making a batch of clips for a single sound, one unreliable input poisons the batch.

If you still want both of you in one shot

If you want to try it, the goal is to remove every ambiguity you can before the upload. That means:

• Both people must be adults, and both must consent.
• Both should be framed the same way: knees-up or full body, ideally side by side.
• Nobody should be half out of frame or cropped in a way the other person is not.
• Both faces should be clear, unobstructed, and lit by the same light source.
• No sunglasses on either person, no hat.
• Stand close enough that you read as one group, not two unrelated people who happen to share a photo.
• Nobody else in the background.

Even with all of that, treat it as an experiment rather than a plan. There is no cheap low-resolution test to run first, because every render costs 30 credits. So decide whether the pair is worth one render before you upload, and remember that a render that fails on our side is refunded automatically.

If your photo has a pet in it rather than a second person, the situation is different again and worse: the generator is built for human bodies, and the pet article explains exactly why a human walk on a pet photo is unpredictable.

The two-clip route

This is the approach we actually recommend for pairs, and it produces the more polished result.

Generate each person separately, from their own single-subject photo. Each clip then gets the full attention of the model: one body, one walk, one stable face, twelve seconds of vertical footage each. Then combine them afterwards in any editing app, or simply post them as a pair of related clips.

A few reasons this works better:

• Both renders are reliable, because both inputs are the format the tool expects.
• You control the pairing in the edit, choosing which clip goes first and how they cut together.
• If one render comes out badly, you redo one clip instead of the whole pair.
• You can still make them feel connected: same sound, matching direction prompts, or a caption that ties the two posts together.

If you are posting the pair, keep the sound consistent across both so the two clips read as a set. The posting workflow is the same as any single clip, and the step-by-step version lives in How to make a Fergie Baby AI video.

Consent rules

Whenever a second person appears in a clip you generate, these are non-negotiable.

Use your own photo, or a friend's photo with their clear consent. Ask them in a way that gets an actual answer, and keep that answer. "They probably would not mind" is not consent.

Never upload a stranger. Never upload a child. Never upload a celebrity. All three are out of scope, and none of them can consent in the way this tool requires.

If a friend changes their mind after you have made a clip, do not repost it. Take it down and do not generate more from their photo. The clip is easy to remake; the trust is not.

Also be careful with images pulled from someone else's profile, group chats, or a shared photo album. Being in the same photo does not imply permission to animate it.

What you get

One photo in, one twelve-second vertical MP4 out: 9:16, silent, no watermark. We ship no music, so the sound comes from inside the app you post to, which keeps the audio credited to its owner. Rendering takes a few minutes.

Cost is per clip: The only spec is 1080p · 12 s = 30 credits. There is no other resolution or length. Failed renders are refunded automatically. Two clips means two renders, so budget accordingly. The full list is on /pricing.

Honest limits

One person per photo is the format the tool is built for. A two-person image can render, but the walk is taken from a single-subject reference clip, so the movement across two bodies is not something we can make predictable.

Adults only. Your own photo, or a friend's with their consent. Never a stranger, never a child, never a celebrity.

We are an independent tool. We are not affiliated with the artist, and we are not officially partnered with any platform. We do not guarantee how any clip performs once posted.

FAQ

Will a two-person photo definitely fail?

No, and that is the frustrating part. Some two-person photos render cleanly, others produce sliding bodies, merged silhouettes, or one person walking while the other stays still. Because the outcome depends on the specific poses in your photo, the only honest answer is that it may work and may not. Decide whether it is worth a render before you spend one.

Can I use a group photo and ask for just me to walk?

No. There is no selection step. The model reads the whole image, so a group photo introduces all the ambiguity of a two-person shot multiplied by the number of people in frame. Crop to one person before uploading if you want a predictable result.

Should my friend and I use the same sound for separate clips?

Yes, if you want the two posts to read as one set. Same sound and matching direction prompts make separate clips feel like a deliberate pair. It also keeps the audio credited to its owner, since you pick the sound inside the platform rather than shipping it with the file.

What if my friend is in the photo but does not want to be animated?

Then do not upload that photo. Ask for a solo photo instead, or take one. There is no way to exclude one person from an image you upload, so consent has to happen before the upload, not after. If they are unsure, offer to make a separate clip from their own solo photo so they can see the result for themselves.

Do I pay for both clips if I use the two-clip route?

Yes, each clip is a separate render and each one costs 30 credits. A pair made from two photos is two renders. If one render fails, its credits are returned automatically, so you are not charged for a clip you never received. See /pricing for the exact numbers.