How to Make an AI Fitness Influencer

A step-by-step guide to building a fitness character on makeinfluencers.com and turning them into before/after transformation content.

Fit vs unfit transformation clip from TikTok — This is the kind of video you'll end up with.

Step 0: Know Who You're Talking To

Before you generate anything, decide who this character is for.

Who's the audience? Men 25–40 trying to lose the dad-bod, women starting their first cut, gym newbies, busy professionals.
What kind of character resonates with them? A relatable everyman, an aspirational athlete, a coach-next-door, a "used-to-be-like-you" figure.

Your character is a casting decision. Get this right and every later step gets easier.

Step 1: Create Your Character

Open MakeInfluencers and head to the character creator. You have two starting points:

Pinterest reference — find a person whose look already fits your audience and use them as your base.
Generate from scratch — describe the character in the prompt and let the model build them.

Either way, build the fit version first. This is your "after" — the aspirational version of the character.

Finished fit character portrait on MakeInfluencers — The finished fit character.

Step 2: Generate the "Before" Version

From the fit character, generate a variation that's the opposite end of the spectrum — heavier, tired, balding, defeated. This is your "before."

You now have two looks of the same person: a good version and a bad version. That contrast is the whole engine of the content.

Generating the 'before' variation of the character inside MakeInfluencers — Generating the unfit variation from the same base character.

Step 3: Generate Matching Shots for Each Look

For each look, generate the same scene so the only thing that changes is the body. For this guide we're doing a gym locker-room progress pic.

Keep the framing, lighting, and pose consistent across both versions — that's what sells the transformation.

Put the character into a pose for the start frame of the video by creating a shot.

Pose creation prompt UI on MakeInfluencers — Set up the shot prompt — use the same one for both looks.

Run the same prompt against both looks and you get matched before/after shots:

Why this works: people love watching transformations. The contrast between before and after is what gets them hooked — it implies a story without you having to tell one.

Step 4: Animate with Image-to-Video

Take each generated shot into the image-to-video flow and animate the pose with a simple prompt — small motions work best (subtle breathing, flexing, looking at the mirror, lifting a shirt).

Keep prompts short. Over-prompting at this stage usually makes the motion weird.

Image-to-video screen with progress pic loaded and motion prompt typed in — Image-to-video screen — keep the motion prompt short.

Animated transformation clip — The animated clip — small motions usually look the most natural.

Step 5: Edit and Post

Every generation is saved automatically in your library, so you don't need to re-render anything. Pull the clips into your editor, cut them together (before → after, with a beat or a sound cue at the transition), and post.

The edited transformation, ready to post.

Recap

Pick your audience and character archetype
Generate the fit version
Generate the unfit variation
Generate the same scene for both looks
Animate each with image-to-video
Edit before → after, post

The contrast does the work. Your job is to make the two versions believable as the same person.