Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create AI-generated presentations, documents, webpages, and social posts via Gamma MCP. Use when user asks to create slides, make a presentation, generate a deck, build a document, or create a webpage.
.claude/skills/bilal140202-gamma-presentation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 101% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 67% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 44% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 76% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 55% | 0% |
> Version: 1.1.0 | Updated: 2026-03-03 | Category: Presentations & Docs
NEVER look for GAMMA_API_KEY in .env files. NEVER call Gamma REST API directly with fetch/requests. NEVER search for credentials — there are none to find.
ONLY correct way → call the MCP tool directly:
mcp__claude_ai_Gamma__generate(inputText, format, themeId, ...)Auth is handled transparently by the MCP connector. Works in Claude Code (VSCode) and claude.ai web. If you see "key empty" or "API key not found" — wrong approach. Stop and use the MCP tool.
Creates AI-generated presentations, documents, webpages, and social posts via Gamma MCP. No API key needed. No setup. Fully live via MCP.
Use for:
Say any of these:
| Tool | Use When | |------|----------| | generate | Create new presentation/doc/webpage/social post | | get_themes | Browse all 100+ available themes | | get_folders | List saved Gamma folders | | get_generation_status | Check if generation is complete |
| Format | Use When | |--------|----------| | presentation | Slides (default — use for most things) | | document | Long-form doc, guide, course notes | | webpage | Single-page website / landing page | | social | LinkedIn carousel / social post |
| Theme ID | Name | Best For | |----------|------|----------| | nebulae | Nebulae | Default brand theme — dark, purple/blue, futuristic, high-tech | | stratos | Stratos | Investor decks, serious tech — deep navy, bold, dramatic | | aurora | Aurora | Creative content, YouTube — dark gradient, fuchsia to purple | | gamma-dark | Gamma Dark | Bold content — dark purple + salmon, playful futuristic | | founder | Founder | Minimalist dark — startup pitches, professional | | velvet-tides | Velvet Tides | Luxury/agency feel — dark black + purple gradient | | default-dark | Basic Dark | Clean professional — dark, blue/purple, simple |
Always default to nebulae unless user asks for something different.
format: presentation | document | webpage | social
themeId: nebulae (default)
numCards: 6-10 (presentations), omit for docs/webpages
textMode: preserve (use exact content) | generate (expand from outline) | condense (shorten long content)
textOptions.amount: brief | medium | detailed | extensive
textOptions.tone: bold, futuristic, professional
textOptions.language: en (default)
exportAs: pptx | pdf (only when user asks)
# Dimensions (cardOptions.dimensions) — 9 options:
16x9 → YouTube, standard widescreen (default)
4x3 → Classic slides
fluid → Responsive, no fixed ratio
letter → US Letter (8.5x11) for print
a4 → A4 for print/PDF
pageless → Infinite scroll document
1x1 → Square (Instagram)
4x5 → Portrait (Instagram feed)
9x16 → Vertical (Instagram Reels, TikTok, Shorts)
# AI Image Models (imageOptions.model) — use for better visuals:
flux-1-quick → Fast, good quality (default)
flux-kontext-fast → Context-aware, faster
flux-kontext-max → Context-aware, max quality
flux-1-pro → Pro quality
imagen-3-pro → Google Imagen 3 Pro (photorealistic)
imagen-4-pro → Google Imagen 4 Pro (best photorealism)
ideogram-v3 → Best for text-in-image
leonardo-phoenix → Artistic/creative style
recraft-v3 → Vector/design style
recraft-v3-svg → SVG output
# Sharing (sharingOptions) — set on generation:
workspaceAccess: edit | comment | view | noAccess
externalAccess: edit | comment | view | noAccess
emailOptions: { recipients: [...], access: 'view' }
# Header/Footer (cardOptions.headerFooter):
Add slide numbers, logo, or text to every slide
topLeft/topRight/bottomLeft/bottomCenter/bottomRight
type: cardNumber | image | textnebulae. Use get_themes if user wants something specific.inputText (use preserve mode for exact content)generate → get generationIdget_generation_status every 5-10s until status: completedgammaUrl with userSlide 1 — [Title Slide]
Title: [Main title]
Subtitle: [Subtitle or tagline]
Tagline: [Optional quote]
Slide 2 — [Section Name]
[Bullet points or paragraph content]
Slide 3 — [Section Name]
[Content...]"Make a 6-slide Gamma presentation about MCP servers for my Euron class"
→ format: presentation, themeId: nebulae, numCards: 6"Create a QuotaHit pitch deck for investors — 10 slides, professional theme"
→ format: presentation, themeId: stratos, numCards: 10"Create a Gamma document with all the class notes from today's session"
→ format: document, themeId: nebulae, textMode: preserve"Create a presentation and export as PPTX"
→ exportAs: pptx| Input | Type | Required | Description | |-------|------|----------|-------------| | inputText | string | Yes | Full content (slides, outline, or paste) | | format | enum | No | presentation / document / webpage / social | | themeId | string | No | Default: nebulae | | numCards | number | No | Number of slides | | textMode | enum | No | preserve / generate / condense | | exportAs | enum | No | pptx / pdf |
| Output | Description | |--------|-------------| | gammaUrl | Shareable link to the finished presentation | | generationId | ID to check status | | credits.remaining | How many Gamma credits left |
| Credential | Where | Notes | |------------|-------|-------| | Gamma MCP | Already connected via Claude MCP | No setup needed |
thumbnail-generator → Create image prompt after building slidesimage-to-video → Animate slides into short videoghost-browser → Auto-post deck to LinkedInsend-telegram → Notify when deck is ready| Source | How | Use When | |--------|-----|----------| | YouTube video | Fetch transcript → feed as inputText | Turn a class recording into a deck | | Web article | Extract article text → feed as inputText + textMode: condense | Research-to-deck | | n8n workflow | Trigger Gamma generation from n8n HTTP node | Auto-generate weekly reports | | Raw text/notes | Paste directly + textMode: preserve | Class notes → student deck | | CSV/data | Summarize data → narrative slides | Sales report decks |
| Use Case | Dimension | |----------|-----------| | YouTube thumbnail / standard deck | 16x9 | | Instagram carousel | 4x5 | | Instagram/TikTok/Shorts | 9x16 | | Print handout | letter or a4 | | LinkedIn document | fluid | | Instagram square | 1x1 |
| Use Case | Model | |----------|-------| | Fast deck (default) | flux-1-quick | | Tech/realistic visuals | imagen-4-pro | | Slides with text in images | ideogram-v3 | | Creative/artistic | leonardo-phoenix | | Design/brand assets | recraft-v3 |
| Sub-Skill | File | Use When | |-----------|------|----------| | Social Media Presentations | See Social-Media-Agent-1.0/ | LinkedIn carousels, YouTube slides, content decks | | Onsite Sales Presentations | See Onsite/presentations/ | Client decks, sales proposals, onboarding slides |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→fail | 4,408 | 4,402 | -0% | 1 | 1 | 0% | 686 | 3,082 | +349% | 0 | 0 | — |
case-03 | fail→fail | 31,858 | 24,347 | -24% | 1 | 1 | 0% | 6,191 | 4,891 | -21% | 0 | 0 | — |
case-04 | fail→pass | 9,401 | 7,509 | -20% | 1 | 1 | 0% | 1,872 | 3,757 | +101% | 0 | 0 | — |
case-01 | fail→fail | 14,259 | 20,237 | +42% | 1 | 1 | 0% | 2,784 | 3,253 | +17% | 0 | 0 | — |
case-05 | pass→pass | 4,842 | 4,753 | -2% | 1 | 1 | 0% | 905 | 3,094 | +242% | 0 | 0 | — |
case-06 | pass→pass | 8,284 | 5,763 | -30% | 1 | 1 | 0% | 1,486 | 3,253 | +119% | 0 | 0 | — |
case-07 | pass→pass | 9,070 | 4,695 | -48% | 1 | 1 | 0% | 1,450 | 3,089 | +113% | 0 | 0 | — |
case-08 | pass→pass | 7,472 | 2,815 | -62% | 1 | 1 | 0% | 1,223 | 2,748 | +125% | 0 | 0 | — |
case-09 | fail→pass | 10,827 | 3,206 | -70% | 1 | 1 | 0% | 1,727 | 2,882 | +67% | 0 | 0 | — |
case-10 | fail→pass | 12,284 | 2,957 | -76% | 1 | 1 | 0% | 1,915 | 2,764 | +44% | 0 | 0 | — |
case-11 | fail→pass | 10,639 | 4,988 | -53% | 1 | 1 | 0% | 1,753 | 3,080 | +76% | 0 | 0 | — |
case-12 | fail→fail | 30,259 | 30,380 | +0% | 1 | 1 | 0% | 6,177 | 3,500 | -43% | 0 | 0 | — |
case-13 | fail→pass | 11,674 | 3,577 | -69% | 1 | 1 | 0% | 1,858 | 2,876 | +55% | 0 | 0 | — |
case-14 | fail→fail | 23,920 | 15,156 | -37% | 1 | 1 | 0% | 4,243 | 3,896 | -8% | 0 | 0 | — |
case-15 | pass→pass | 6,129 | 3,185 | -48% | 1 | 1 | 0% | 954 | 2,844 | +198% | 0 | 0 | — |
case-16 | fail→pass | 8,677 | 3,506 | -60% | 1 | 1 | 0% | 1,436 | 2,868 | +100% | 0 | 0 | — |
case-17 | fail→fail | 22,616 | 19,911 | -12% | 1 | 1 | 0% | 4,608 | 3,494 | -24% | 0 | 0 | — |
case-18 | pass→pass | 7,800 | 2,455 | -69% | 1 | 1 | 0% | 1,240 | 2,688 | +117% | 0 | 0 | — |
case-19 | pass→pass | 3,728 | 2,808 | -25% | 1 | 1 | 0% | 615 | 2,756 | +348% | 0 | 0 | — |
case-20 | fail→pass | 8,670 | 2,633 | -70% | 1 | 1 | 0% | 1,407 | 2,769 | +97% | 0 | 0 | — |
case-21 | pass→pass | 5,287 | 2,378 | -55% | 1 | 1 | 0% | 819 | 2,703 | +230% | 0 | 0 | — |
case-22 | fail→pass | 9,094 | 3,321 | -63% | 1 | 1 | 0% | 1,366 | 2,933 | +115% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 18 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +36 percentage points is the difference between those two pass rates over the 18 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.