guides/social/social-media-carousel/SKILL.md
Multi-slide carousel design for Instagram, LinkedIn, and Twitter/X with layout rules and hooks. Covers slide structure, text hierarchy, swipe psychology, and platform-specific specs. Use for: carousel posts, Instagram carousels, LinkedIn carousels, slide posts, educational content. Triggers: carousel, instagram carousel, linkedin carousel, slide post, carousel design, swipe post, multi-image post, carousel template, educational carousel, carousel content, instagram slides, linkedin slides
npx skillsauth add inf-sh/skills social-media-carouselInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security scan pending...
This skill is queued for security scanning. Results will appear when the scan completes.
Install the belt CLI skill:
npx skills add belt-sh/cli
Design high-engagement carousel posts via inference.sh CLI.
Requires inference.sh CLI (
belt). Install instructions
belt login
# Generate a carousel slide
belt app run infsh/html-to-image --input '{
"html": "<div style=\"width:1080px;height:1080px;background:#0f172a;display:flex;align-items:center;justify-content:center;padding:80px;font-family:system-ui;color:white;text-align:center\"><div><p style=\"font-size:24px;color:#818cf8;text-transform:uppercase;letter-spacing:3px\">5 Rules for</p><h1 style=\"font-size:64px;margin:16px 0;font-weight:900;line-height:1.1\">Writing Headlines That Convert</h1><p style=\"font-size:22px;opacity:0.5;margin-top:24px\">Swipe →</p></div></div>"
}'
| Platform | Dimensions | Slides | Aspect Ratios | |----------|-----------|--------|---------------| | Instagram | 1080 x 1080 px | Up to 20 | 1:1 (default), 4:5, 16:9 | | LinkedIn | 1080 x 1080 px or 1080 x 1350 | Up to 20 | 1:1, 4:5 | | Twitter/X | 1080 x 1080 px | Up to 4 | 1:1, 16:9 | | Facebook | 1080 x 1080 px | Up to 10 | 1:1, 4:5 |
Use 1080 x 1350 (4:5) on Instagram and LinkedIn — takes up more screen real estate in the feed than square.
| Slide | Purpose | Content | |-------|---------|---------| | 1 | Hook | Bold claim, question, or promise — stops the scroll | | 2 | Context | Why this matters, set up the problem | | 3-6 | Value | One point per slide, numbered | | 7 | CTA | Follow, save, share, comment, visit link |
The most important slide. If this fails, nobody swipes.
| Hook Type | Example | |-----------|---------| | Bold claim | "90% of landing pages make this mistake" | | Question | "Why do your ads get clicks but no conversions?" | | Number + promise | "7 Python tricks I wish I learned sooner" | | Contrarian | "Stop writing blog posts (do this instead)" | | Before/after | Show transformation |
# Hook slide
belt app run infsh/html-to-image --input '{
"html": "<div style=\"width:1080px;height:1350px;background:linear-gradient(180deg,#1e1b4b,#312e81);display:flex;align-items:center;justify-content:center;padding:80px;font-family:system-ui;color:white;text-align:center\"><div><h1 style=\"font-size:72px;font-weight:900;line-height:1.15;margin:0\">90% of Landing Pages Make This Mistake</h1><p style=\"font-size:28px;opacity:0.6;margin-top:32px\">Swipe to find out →</p></div></div>"
}'
One point per slide. Never cram multiple ideas.
# Content slide template
belt app run infsh/html-to-image --input '{
"html": "<div style=\"width:1080px;height:1350px;background:#1e1b4b;padding:80px;font-family:system-ui;color:white;display:flex;flex-direction:column;justify-content:center\"><div><p style=\"font-size:120px;font-weight:900;color:#818cf8;margin:0;line-height:1\">01</p><h2 style=\"font-size:48px;margin:24px 0 16px;font-weight:800;line-height:1.2\">Your headline is too vague</h2><p style=\"font-size:26px;opacity:0.8;line-height:1.6\">\"Welcome to our platform\" tells the visitor nothing. Lead with the outcome: \"Ship docs in minutes, not days.\"</p></div></div>"
}'
# CTA slide
belt app run infsh/html-to-image --input '{
"html": "<div style=\"width:1080px;height:1350px;background:linear-gradient(180deg,#312e81,#1e1b4b);display:flex;align-items:center;justify-content:center;padding:80px;font-family:system-ui;color:white;text-align:center\"><div><h2 style=\"font-size:56px;font-weight:900;margin:0;line-height:1.2\">Found this useful?</h2><p style=\"font-size:32px;opacity:0.8;margin-top:24px;line-height:1.5\">Save this post for later 🔖<br>Follow for more tips</p><p style=\"font-size:24px;opacity:0.4;margin-top:40px\">@yourusername</p></div></div>"
}'
| Element | Size (at 1080px) | Weight | |---------|-----------------|--------| | Slide number | 96-120px | Black (900) | | Heading | 48-64px | Bold (700-800) | | Body text | 24-28px | Regular (400) | | Caption/tag | 18-22px | Medium (500) |
| Rule | Value | |------|-------| | Max words per slide | 30-40 | | Max lines of body text | 4-5 | | Line height | 1.5-1.6 | | Font | Sans-serif (Inter, Montserrat, Poppins) | | Text contrast | 4.5:1 minimum (WCAG AA) |
| Element | Keep Consistent Across All Slides | |---------|----------------------------------| | Background color/gradient | Same palette, slight variations OK | | Font family | Same font throughout | | Text alignment | Same position (left or center) | | Margins/padding | Same spacing | | Accent color | Same highlight color | | Numbering style | Same format (01, 02 or 1., 2.) |
Slide 1: "5 CSS tricks you need to know"
Slide 2: Trick 1 with code example
Slide 3: Trick 2 with code example
...
Slide 6: Trick 5 with code example
Slide 7: "Follow for more dev tips"
Slide 1: "How we grew from 0 to $1M ARR"
Slide 2: The beginning (context)
Slide 3: The challenge
Slide 4: What we tried (failed)
Slide 5: What worked
Slide 6: The result (numbers)
Slide 7: Key takeaway + CTA
Slide 1: "I redesigned this landing page"
Slide 2: Before screenshot
Slide 3: Problem 1 annotated
Slide 4: After screenshot
Slide 5: Improvement 1 explained
Slide 6: Results (conversion lift)
Slide 7: "Want a review? DM me"
Slide 1: "10 tools every designer needs in 2025"
Slides 2-6: 2 tools per slide with logo + one-liner
Slide 7: "Save this for later 🔖"
| Principle | Application | |-----------|------------| | Curiosity gap | Hook promises value that requires swiping | | Numbered progress | "3/7" creates completion drive | | Visual continuity | Consistent design signals "there's more" | | Increasing value | Best tip last — rewards completing | | Swipe cue | Arrow or "Swipe →" on slide 1 |
# Generate all slides for a carousel
for i in 1 2 3 4 5 6 7; do
belt app run infsh/html-to-image --input "{
\"html\": \"<div style='width:1080px;height:1350px;background:#1e1b4b;display:flex;align-items:center;justify-content:center;padding:80px;font-family:system-ui;color:white'><div style='text-align:center'><p style='font-size:28px;opacity:0.5'>Slide $i of 7</p></div></div>\"
}" --no-wait
done
# Generate illustrations for each slide
belt app run falai/flux-dev-lora --input '{
"prompt": "minimal flat illustration, person at desk with laptop, clean modern style, simple shapes, limited color palette purple and blue tones, white background, icon style",
"width": 1080,
"height": 1080
}'
| Mistake | Problem | Fix | |---------|---------|-----| | Weak hook (slide 1) | Nobody swipes | Bold claim, question, or number + promise | | Too much text per slide | Overwhelming, stops reading | Max 30-40 words per slide | | No visual consistency | Looks like different posts | Same colors, fonts, margins throughout | | No swipe indicator | People don't realize there's more | Add "Swipe →" or arrow on slide 1 | | No CTA on last slide | Missed engagement opportunity | Ask to save, follow, share, or comment | | Inconsistent numbering | Feels disorganized | Same number format on every content slide | | Cramming 2+ ideas per slide | Hard to digest | One point per slide, always | | Square format on Instagram | Wastes feed real estate | Use 1080x1350 (4:5) for more visibility |
npx skills add inference-sh/skills@ai-image-generation
npx skills add inference-sh/skills@content-repurposing
npx skills add inference-sh/skills@linkedin-content
Browse all apps: belt app store
data-ai
Generate multi-person talking head podcast videos from scratch using AI — character creation, TTS, avatar animation, and video stitching. Use when the user wants to create a podcast, talking head video, or multi-speaker conversation video.
tools
Generate videos with ByteDance Seedance 2.0 via inference.sh CLI. Unified model for text-to-video, image-to-video, and reference-to-video with synchronized audio, up to 1080p, 4-15s duration. Pro and Fast variants. Studio variants with private asset library for portrait consistency. Use for: social media videos, music videos, product demos, animated content, AI video with sound. Triggers: seedance, seedance 2, bytedance video, seedance t2v, seedance i2v, seedance r2v, video with audio, seedance 2.0, bytedance seedance, seedance studio
tools
Generate talking head avatar videos with Pruna P-Video-Avatar via inference.sh CLI. Turn a portrait image into a realistic speaking video with built-in TTS. 18x faster and 6x cheaper than competitors. Models: P-Video-Avatar, P-Image (for portrait generation). Capabilities: text-to-avatar, audio-driven avatars, 30 voices, 10 languages, 720p/1080p, built-in TTS, dynamic backgrounds, full-body control. Use for: AI presenters, product demos, explainer videos, virtual influencers, marketing, education, multilingual content, UGC, gaming avatars. Triggers: avatar video, talking head, ai avatar, p-video-avatar, pruna avatar, video avatar, ai presenter, digital human, virtual presenter, lipsync, talking avatar, ai spokesperson, heygen alternative, synthesia alternative, veed alternative, fabric alternative, omnihuman alternative
tools
Generate and edit videos with Alibaba HappyHorse 1.0 models via inference.sh CLI. Models: HappyHorse T2V, I2V, R2V, Video Edit. Capabilities: text-to-video, image-to-video, reference-to-video, video editing with natural language, character preservation, 720P/1080P, up to 15 seconds. Use for: physically realistic video, video editing, character-consistent content, product demos, social media. Triggers: happyhorse, happy horse, alibaba video, happyhorse 1.0, dashscope video, alibaba happyhorse, video editing ai, ai video editor