NEW Advanced AI Video Inpainting with Runway’s Act 2 AI Feature
This workflow unlocks something new - and it’s shockingly simple…
You don’t need to swap your whole face or body anymore.
Now you can change just *one part* of the subject - while keeping the rest totally real and in sync.
Works for NON HUMAN CHARACTERS TOO!!
In this test:
- I kept my *real head*, and swapped in robotic arms + chest
- Then flipped it - kept my body, but swapped in a dog face
All driven by one single video performance.
Here’s what makes it possible:
- Runway Act 2 tries to stay loyal to your input image
- So if that image is *partially edited*, it will preserve the rest
- That gives you frame-level control over what gets transformed
To do it:
1. Record a short driver video (camera still, clear action)
2. Grab the *first frame*
3. Use any tool (Supercreator, Photoshop, Ideogram, or even ChatGPT Image Editor)
to modify **just the part** you want to change
4. Feed both the original video + edited image into Act 2
5. Adjust gesture + expression settings to fine-tune realism
The result?
Instant VFX-style character edits - no rotoscoping, no compositing, no re-recording.
You can:
- Keep facial performance, change outfit
- Add a creature mask while preserving body motion
- Stylize just one feature for subtle surreal effects
- Create hybrid personas for storytelling or comedy
It’s perfect for creators who want more realism and more control - not just filters.
The best part? This works *right now*.
You can try it with any image editing tool, any concept, and remix scenes in endless ways.
MASSIVE for anyone wanting to make creative, original content EASILY
PS: I’m testing workflows like this daily - follow for more 🔥
Видео NEW Advanced AI Video Inpainting with Runway’s Act 2 AI Feature канала TechGuyver - Always be creating
You don’t need to swap your whole face or body anymore.
Now you can change just *one part* of the subject - while keeping the rest totally real and in sync.
Works for NON HUMAN CHARACTERS TOO!!
In this test:
- I kept my *real head*, and swapped in robotic arms + chest
- Then flipped it - kept my body, but swapped in a dog face
All driven by one single video performance.
Here’s what makes it possible:
- Runway Act 2 tries to stay loyal to your input image
- So if that image is *partially edited*, it will preserve the rest
- That gives you frame-level control over what gets transformed
To do it:
1. Record a short driver video (camera still, clear action)
2. Grab the *first frame*
3. Use any tool (Supercreator, Photoshop, Ideogram, or even ChatGPT Image Editor)
to modify **just the part** you want to change
4. Feed both the original video + edited image into Act 2
5. Adjust gesture + expression settings to fine-tune realism
The result?
Instant VFX-style character edits - no rotoscoping, no compositing, no re-recording.
You can:
- Keep facial performance, change outfit
- Add a creature mask while preserving body motion
- Stylize just one feature for subtle surreal effects
- Create hybrid personas for storytelling or comedy
It’s perfect for creators who want more realism and more control - not just filters.
The best part? This works *right now*.
You can try it with any image editing tool, any concept, and remix scenes in endless ways.
MASSIVE for anyone wanting to make creative, original content EASILY
PS: I’m testing workflows like this daily - follow for more 🔥
Видео NEW Advanced AI Video Inpainting with Runway’s Act 2 AI Feature канала TechGuyver - Always be creating
Комментарии отсутствуют
Информация о видео
19 ч. 6 мин. назад
00:00:11
Другие видео канала