Steal My Carousels
Two Claude Skills and one tool. Free.
The skill that writes your content
- Download
carousel-builder.zipbelow - Claude: Settings → Capabilities → Skills → "Upload skill" → select the .zip, no need to unzip it
/carousel-builder — I want a carousel on [topic].
The skill that writes your image prompts
Once the copy is locked, say "give me the prompt for slide 1" — this second skill takes over and turns your carousel copy into a ChatGPT-ready image prompt. Style is set on slide 1; every slide after matches automatically.
- Download
chatgpt-carousel-prompt-writer.zipbelow (same install steps as above) - Say the prompt below — it gives back a ChatGPT-ready image prompt
- Paste that into ChatGPT and generate. Repeat per slide — no Canva, no manual design
Give me the ChatGPT image prompt for slide 1.
The tool that removes the AI tag — save this
Platforms read hidden metadata and auto-label AI images. Strip it, locally in your browser, before you upload.
bytemint.ai/clean
Post it — straight from the folder
That's the whole pipeline: Claude skill → ChatGPT → clean the tags → post. No Canva, no editing app in between.
What You Get
Content, structured
Cover line under 8 words. A confirm slide that opens the real problem. One idea per slide. One slide flagged save-worthy. A caption with new context and a single CTA.
Visuals, locked to a system
One ChatGPT-ready prompt per slide. Style locked on slide 1, every slide after matches automatically. Copy-paste ready, no manual design work.
Clean before you post
Every AI-generated image carries hidden metadata — C2PA, IPTC, EXIF, XMP — that platforms read to label content. bytemint.ai/clean strips it locally, in your browser, before you upload.
Two-minute setup
Download the .zip, upload it in Claude's Skills settings. No coding, no config.
All Skills In This Bundle
-
carousel-builderWrites the full carousel — cover line, slide-by-slide copy, one save-worthy slide, caption, and CTA — in a locked voice and structure.
-
chatgpt-carousel-prompt-writerTurns locked carousel copy into ChatGPT-ready image generation prompts, one slide at a time, with visual consistency enforced across the whole set.
