First, look at this 👇
**
This is made by AI.
Moreover, it can be created by combining the currently trending Muse.
Since I've earned 30 million yen with AI videos,
I'm giving away my thinking and know-how for free.
Please grab that too 👇
▶︎ Newsletter Dumping Ground for Thoughts
Now, let's get to the main topic.
1. Register for Muse
https://muse.ai/?aymh_complete=1
Muse is an AI that creates character settings and images
just by asking in conversation.
Open Muse from the Meta AI app or website.
Log in with your Meta account.
The screen is a normal chat.
Just ask in Japanese here.
2. Decide on the Character in Muse
Let's create the content first.
Name, age, vibe, speaking style, clothing preferences.
If these are decided, you can generate
post captions and fan club content all from this setting.
Send this 👇
1Create a character setting for a girl who posts dance videos on SNS.2- Name: [Example: 3 easy-to-call name ideas]3- Age setting: Adult in her 20s. Clearly looks adult.4- Vibe: [Example: Bright, laid-back, slightly airheaded]5- Speaking style: [Example: Soft sentence endings, common catchphrases]6- Clothing taste: [Example: Street style, sporty]7- Likes and dislikes8- Short self-introduction for profile (include a line indicating it's an AI-created character)9Conditions: Do not resemble real people, celebrities, or existing characters. Keep it all-ages.10Once decided, please remember this character setting.
Just rewrite the parts inside [ ] as you like.
Honestly, even if you leave them blank,
Muse will think of something there.
The important part is the last sentence.
When you ask "remember," Muse will retain
this character's settings.
So next time, just saying
"with the remembered character" works.
3. Create the Baseline Image in Muse
Next is the one image that becomes
this character's "correct face."
Front-facing face and full body from head to toe.
Plain background, vertical format.
From now on, you'll always attach
and use this one image 👇
1Create one baseline image for the character you remembered.2- Vertical format3- Plain light gray background4- Left: Close-up of front-facing face; Right: Full body from head to toe (standing straight, facing forward)5- Outfit: [Example: White hoodie, wide pants, sneakers]6- Expression: [Example: Soft smile]7- Must be an adult woman in her 20s, clearly looking adult8- Do not resemble real people or celebrities9- No text included. If you don't like the hair or clothes,10you can fix it via conversation like "make bangs a bit shorter."
Don't use photos of others or faces of real people;
start with a face created from scratch.
Save it once decided.
It will appear many times from here on.
4. Change Outfits and Locations in Muse
Same clothes, same background every day.
That gets boring quickly lol
So, attach the baseline image and
create several images changing only the outfit and location 👇
1Using the character from the attached image, create images changing only the outfit and location.2- Keep face, hairstyle, and body type exactly as in the attached image3- Outfit: [Example: Denim jacket, long skirt, short boots]4- Location: [Example: Rooftop at sunset]5- Vertical 9:16, full body visible from head to toe6- All-ages appropriate regular clothes7- No text included. Rewrite outfits and locations,8creating 3-4 images makes it easier.
Cafes, beaches, night streets, your room.
Scattering locations makes it feel like
"every day is different" when turned into videos.
5. Register for Kling

Here, pass the baton to video AI.
While Muse can make videos, Kling is stronger
for making characters dance properly.
Register here 👇
You can register with a Google account or email.
There are free credits, so try those first.
6. Make Images Dance in Kling
Use the "Image to Video" feature.
Input one image created in step 4, and write this in the prompt field 👇
1A young adult woman dances with simple, slow choreography: swaying side to side, small steps, waving her hands, one gentle turn.2Keep her face, hairstyle, and outfit exactly the same as the image.3Fixed camera, full body visible from head to toe, vertical 9:16.4Location: [Example: rooftop at sunset]5Smooth natural movement, family-friendly.
The content is simply:
"Dance slowly. Keep face and clothes as in the image.
Fixed camera, full body, vertical."
Two tips 👇
✅ Choreography should be slow
✅ Camera should not move
Slow movements keep faces and hands clearer.
Swap images and create multiple videos
with different locations.
For more advanced dancing, use motion control.
If you input a video where someone already danced
as a reference, you can copy that movement
to the character directly.
Faces don't need to be visible.
Full body inclusion is sufficient.
Only use videos you filmed yourself as references.
Do not use other people's dance videos.
7. Connect and Add BGM in CapCut

Arrange the finished videos in CapCut
and combine them into one.
Just connecting videos from different locations
creates a feeling of dancing everywhere.
Add BGM in CapCut, or choose TikTok music
when uploading to TikTok.
Short length is sufficient.
8. Post to TikTok

Create an account for the character and
upload videos and captions.
Pace: 1 video per day.
The reason you created multiple images in step 4
is to avoid running out of content.
When posting, turn on
the AI-generated content label.
9. Register for Fanvue

Once followers grow, prepare the fan club.
Fanvue is a fan club site
where supporters pay monthly.
You can create accounts even for AI characters.
Register here 👇
During registration, identity verification
of the operator (you) is required.
Write in the profile that it is
an AI character.
10. Receive Sales
Fanvue's split is 80% for creators.
Accumulated sales can be withdrawn
to bank accounts etc.
Know-how like this AI video workflow
is often distributed guerrilla-style
in my newsletter below.
▶︎ Newsletter Dumping Ground for Thoughts
Registration is free, so sign up early!!





