AI-Character-Agent-Kit.zip
10.9 KB
Built the agent that runs my entire AI character pipeline and I'm giving it away.
Drop it into Claude Code, give it a niche, and one message later you have an original character, a face prompt, a pinned manifesto, and 30 days of content, every Nanobanana Pro frame and every Kling clip prompt already written.
No manual setup. No re-explaining the character between sessions. It reads its own state and picks up where it left off.
Drop it into Claude Code, give it a niche, and one message later you have an original character, a face prompt, a pinned manifesto, and 30 days of content, every Nanobanana Pro frame and every Kling clip prompt already written.
No manual setup. No re-explaining the character between sessions. It reads its own state and picks up where it left off.
❤4
AI-Character-Agent-Kit.zip
74.9 KB
use this file in claude code
the skills weren't properly uploaded in the old file
this is the updated version
the skills weren't properly uploaded in the old file
this is the updated version
Paper_Yapper_Ad_Workflow_Example.pdf
590.8 KB
i said i'd post the exact prompts. wrote up the entire workflow instead
script, character lock, clip planning, every frame prompt, every video prompt, the exact failure points and how to fix them
FULL BREAKDOWN
script, character lock, clip planning, every frame prompt, every video prompt, the exact failure points and how to fix them
FULL BREAKDOWN
This media is not supported in your browser
VIEW IN TELEGRAM
here's the guide 👇
everything's in there: character lock, frame prompts, kling clips, banner
and particle overlays, all the exact prompts, nothing held back.
start with the anchor image and clip 1 before running the whole sequence,
that's the fastest way to know if it's gonna work for your product.
lmk if you end up using it, would love to see what you build.
everything's in there: character lock, frame prompts, kling clips, banner
and particle overlays, all the exact prompts, nothing held back.
start with the anchor image and clip 1 before running the whole sequence,
that's the fastest way to know if it's gonna work for your product.
lmk if you end up using it, would love to see what you build.