guides/design/pitch-deck-visuals/SKILL.md
Investor pitch deck structure with slide-by-slide framework, visual design rules, and data presentation. Covers the 12-slide framework, chart types, team slides, and common investor turn-offs. Use for: fundraising decks, investor presentations, startup pitch, demo day, grant proposals. Triggers: pitch deck, investor deck, startup pitch, fundraising deck, demo day, pitch presentation, investor presentation, seed deck, series a deck, pitch slides, startup presentation, vc pitch, investor meeting
npx skillsauth add inference-sh-0/skills pitch-deck-visualsInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Install the belt CLI skill:
npx skills add belt-sh/cli
Create investor-ready pitch deck visuals via inference.sh CLI.
Requires inference.sh CLI (
belt). Install instructions
belt login
# Generate a slide background
belt app run infsh/html-to-image --input '{
"html": "<div style=\"width:1920px;height:1080px;background:linear-gradient(135deg,#0f0f23,#1a1a3e);display:flex;align-items:center;padding:100px;font-family:system-ui;color:white\"><div><p style=\"font-size:24px;color:#818cf8;text-transform:uppercase;letter-spacing:3px\">The Problem</p><h1 style=\"font-size:72px;margin:16px 0;font-weight:800;line-height:1.1\">Teams waste 12 hours/week on manual reporting</h1><p style=\"font-size:28px;opacity:0.7\">Source: Forrester Research, 2024</p></div></div>"
}'
| # | Slide | Duration | Content | |---|-------|----------|---------| | 1 | Title | 15s | Company name, tagline, your name | | 2 | Problem | 45s | Pain point with data | | 3 | Solution | 45s | Your product in one sentence | | 4 | Demo/Product | 60s | Screenshot or live demo | | 5 | Market Size | 30s | TAM → SAM → SOM | | 6 | Business Model | 30s | How you make money | | 7 | Traction | 45s | Growth metrics, customers | | 8 | Competition | 30s | Positioning, not feature list | | 9 | Team | 30s | Why you specifically will win | | 10 | Financials | 30s | Revenue projections, unit economics | | 11 | The Ask | 15s | How much, what for | | 12 | Contact | 10s | Email, next steps |
Total: ~6 minutes. Never exceed 20 slides.
| Element | Size (1920x1080) | Rule | |---------|-----------------|------| | Slide title | 48-72px | Max 6 words | | Key stat/number | 96-144px | One per slide, when applicable | | Body text | 24-32px | Max 6 bullet points | | Caption/source | 16-20px | Cite data sources | | Font | Sans-serif only | Inter, Helvetica, SF Pro, or similar |
If you need more text, you need more slides.
| Element | Guideline | |---------|-----------| | Background | Dark (navy, charcoal) OR clean white — pick one, commit | | Accent color | ONE brand color for emphasis | | Text | White on dark, or dark grey (#333) on light | | Charts | 2-3 colors max, your brand color = "you" | | Avoid | Gradients on text, neon colors, more than 3 colors |
| Rule | Why | |------|-----| | Consistent margins (80-100px) | Professional, clean | | Left-align text (never center body text) | Easier to scan | | One visual per slide | Focus attention | | Slide numbers | Helps investors reference specific slides | | Logo in corner | Subtle brand reinforcement |
belt app run infsh/html-to-image --input '{
"html": "<div style=\"width:1920px;height:1080px;background:#0f0f23;display:flex;align-items:center;justify-content:center;font-family:system-ui;color:white;text-align:center\"><div><h1 style=\"font-size:80px;font-weight:900;margin:0\">DataFlow</h1><p style=\"font-size:32px;opacity:0.7;margin-top:16px\">Automated reporting for data teams</p><p style=\"font-size:22px;opacity:0.5;margin-top:40px\">Seed Round — Q1 2025</p></div></div>"
}'
One big number + one sentence.
belt app run infsh/html-to-image --input '{
"html": "<div style=\"width:1920px;height:1080px;background:#0f0f23;display:flex;align-items:center;padding:100px;font-family:system-ui;color:white\"><div><p style=\"font-size:24px;color:#f59e0b;text-transform:uppercase;letter-spacing:3px;margin:0\">The Problem</p><h1 style=\"font-size:144px;margin:20px 0;font-weight:900;color:#f59e0b\">12 hrs/week</h1><p style=\"font-size:36px;opacity:0.8;line-height:1.4\">The average data analyst spends 12 hours per week<br>building reports manually</p><p style=\"font-size:20px;opacity:0.4;margin-top:30px\">Source: Forrester Research, 2024</p></div></div>"
}'
Use concentric circles, not pie charts:
belt app run infsh/python-executor --input '{
"code": "import matplotlib.pyplot as plt\nimport matplotlib\nmatplotlib.use(\"Agg\")\n\nfig, ax = plt.subplots(figsize=(19.2, 10.8))\nfig.patch.set_facecolor(\"#0f0f23\")\nax.set_facecolor(\"#0f0f23\")\n\ncircles = [\n (0, 0, 4.0, \"#1e1e4a\", \"TAM\\n$50B\", 40),\n (0, 0, 2.8, \"#2a2a5a\", \"SAM\\n$8B\", 32),\n (0, 0, 1.4, \"#818cf8\", \"SOM\\n$800M\", 28)\n]\n\nfor x, y, r, color, label, fontsize in circles:\n circle = plt.Circle((x, y), r, color=color, ec=\"#333366\", linewidth=2)\n ax.add_patch(circle)\n ax.text(x, y, label, ha=\"center\", va=\"center\", fontsize=fontsize, color=\"white\", fontweight=\"bold\")\n\nax.set_xlim(-5, 5)\nax.set_ylim(-5, 5)\nax.set_aspect(\"equal\")\nax.axis(\"off\")\nax.text(0, 4.8, \"Market Opportunity\", ha=\"center\", fontsize=36, color=\"white\", fontweight=\"bold\")\nplt.tight_layout()\nplt.savefig(\"market-size.png\", dpi=100, facecolor=\"#0f0f23\")\nprint(\"Saved\")"
}'
Show growth, not just numbers. Up-and-to-the-right chart.
belt app run infsh/python-executor --input '{
"code": "import matplotlib.pyplot as plt\nimport matplotlib\nmatplotlib.use(\"Agg\")\n\nfig, ax = plt.subplots(figsize=(19.2, 10.8))\nfig.patch.set_facecolor(\"#0f0f23\")\nax.set_facecolor(\"#0f0f23\")\n\nmonths = [\"Jan\", \"Feb\", \"Mar\", \"Apr\", \"May\", \"Jun\", \"Jul\", \"Aug\"]\nrevenue = [8, 12, 18, 28, 42, 58, 82, 120]\n\nax.fill_between(range(len(months)), revenue, alpha=0.3, color=\"#818cf8\")\nax.plot(range(len(months)), revenue, color=\"#818cf8\", linewidth=4, marker=\"o\", markersize=10)\nax.set_xticks(range(len(months)))\nax.set_xticklabels(months, color=\"white\", fontsize=18)\nax.tick_params(colors=\"white\", labelsize=16)\nax.set_ylabel(\"MRR ($K)\", color=\"white\", fontsize=20)\nax.spines[\"top\"].set_visible(False)\nax.spines[\"right\"].set_visible(False)\nax.spines[\"bottom\"].set_color(\"#333\")\nax.spines[\"left\"].set_color(\"#333\")\nax.set_title(\"Monthly Recurring Revenue\", color=\"white\", fontsize=32, fontweight=\"bold\", pad=20)\nax.text(7, 120, \"$120K MRR\", color=\"#22c55e\", fontsize=28, fontweight=\"bold\", ha=\"center\", va=\"bottom\")\nax.text(7, 112, \"15x growth in 8 months\", color=\"#22c55e\", fontsize=18, ha=\"center\")\nplt.tight_layout()\nplt.savefig(\"traction.png\", dpi=100, facecolor=\"#0f0f23\")\nprint(\"Saved\")"
}'
Never use a feature matrix against competitors. Use a 2x2 positioning map.
# See the competitor-teardown skill for positioning map generation
belt app run infsh/python-executor --input '{
"code": "import matplotlib.pyplot as plt\nimport matplotlib\nmatplotlib.use(\"Agg\")\n\nfig, ax = plt.subplots(figsize=(19.2, 10.8))\nfig.patch.set_facecolor(\"#0f0f23\")\nax.set_facecolor(\"#0f0f23\")\n\ncompetitors = {\n \"Us\": (0.6, 0.7, \"#22c55e\", 300),\n \"Legacy Tool\": (-0.5, 0.5, \"#6366f1\", 200),\n \"Startup X\": (0.3, -0.4, \"#6366f1\", 200),\n \"Manual Process\": (-0.6, -0.6, \"#475569\", 150)\n}\n\nfor name, (x, y, color, size) in competitors.items():\n ax.scatter(x, y, s=size*5, c=color, zorder=5, alpha=0.8)\n weight = \"bold\" if name == \"Us\" else \"normal\"\n ax.annotate(name, (x, y), textcoords=\"offset points\", xytext=(15, 15), fontsize=22, color=\"white\", fontweight=weight)\n\nax.axhline(y=0, color=\"#333\", linewidth=1)\nax.axvline(x=0, color=\"#333\", linewidth=1)\nax.set_xlim(-1, 1)\nax.set_ylim(-1, 1)\nax.set_xlabel(\"Manual ← → Automated\", fontsize=22, color=\"white\", labelpad=15)\nax.set_ylabel(\"Basic ← → Advanced\", fontsize=22, color=\"white\", labelpad=15)\nax.set_title(\"Competitive Landscape\", fontsize=32, color=\"white\", fontweight=\"bold\", pad=20)\nax.tick_params(colors=\"#0f0f23\")\nfor spine in ax.spines.values():\n spine.set_visible(False)\nplt.tight_layout()\nplt.savefig(\"competition.png\", dpi=100, facecolor=\"#0f0f23\")\nprint(\"Saved\")"
}'
# Generate professional team headshots/avatars
belt app run falai/flux-dev-lora --input '{
"prompt": "professional headshot portrait, person in business casual attire, clean neutral background, warm studio lighting, confident friendly expression, corporate photography style",
"width": 512,
"height": 512
}'
Layout: Photos in a row with name, title, and one credential each.
| Person | Format | |--------|--------| | CEO | Name, title, "Ex-[Company], [credential]" | | CTO | Name, title, "Built [thing] at [Company]" | | Others | Name, title, one relevant credential |
Max 4 people on the team slide. More = unfocused.
| Chart Type | Use For | Never Use For | |-----------|---------|--------------| | Line chart | Growth over time (traction) | Comparisons between categories | | Bar chart | Comparing amounts | Time series (use line) | | Concentric circles | TAM/SAM/SOM | Anything else | | 2x2 matrix | Competitive positioning | Feature comparison | | Single big number | Key metric highlight | Multiple metrics | | Pie chart | NEVER | Anything (hard to read, unprofessional) |
| Rule | Why | |------|-----| | Max 2 colors per chart | Clarity | | Your company = green or brand color | Positive association | | Label directly on chart | No separate legend needed | | Remove gridlines or make very subtle | Reduce clutter | | Start Y-axis at 0 | Don't mislead | | Cite data sources | Credibility |
| Slide | Investor's Real Question | |-------|------------------------| | Problem | "Is this a real problem people pay to solve?" | | Solution | "Is this 10x better than the status quo?" | | Market | "Is this big enough to matter?" | | Traction | "Is this actually working?" | | Team | "Can these people execute?" | | Ask | "Is this a reasonable deal?" |
| Mistake | Problem | Fix | |---------|---------|-----| | Too many slides (20+) | Loses attention, unfocused | Max 12-15 slides | | Wall of text | Nobody reads it | 1-6-6 rule: 1 idea, 6 words, 6 bullets | | Feature comparison table vs competitors | Looks defensive | Use 2x2 positioning map | | Pie charts | Hard to read, unprofessional | Use bar charts or big numbers | | No data sources cited | Looks made up | Always cite sources | | Team slide with 8+ people | Unfocused | Max 4, focus on relevant experience | | Inconsistent design | Looks amateur | Same colors, fonts, margins on every slide | | No "The Ask" slide | Investor doesn't know what you want | State amount, use of funds, timeline | | Vanity metrics | "1M visits" means nothing without conversion | Show revenue, active users, retention | | Too much product demo | This is a business pitch, not a demo | Max 2 slides on product, focus on business |
npx skills add inference-sh/skills@competitor-teardown
npx skills add inference-sh/skills@data-visualization
npx skills add inference-sh/skills@ai-image-generation
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