v1.0.11

Prezentit

VeGoVeVO VeGoVeVO ← All skills

Generate beautiful AI-powered presentations instantly. Create professional slides with custom themes, visual designs, and speaker notes—all through natural language commands.

Downloads
1.5k
Stars
1
Versions
12
Updated
2026-02-24

Install

npx clawhub@latest install prezentit

Documentation

``skill

---

name: prezentit

description: Generate beautiful AI-powered presentations instantly. Create professional slides with custom themes, visual designs, and speaker notes—all through natural language commands.

homepage: https://prezentit.net

emoji: "👽"

metadata:

clawdbot:

emoji: "👽"

skillKey: prezentit

homepage: https://prezentit.net

requires:

config:

- PREZENTIT_API_KEY

config:

requiredEnv:

- name: PREZENTIT_API_KEY

description: Your Prezentit API key (starts with pk_). Get one free at https://prezentit.net/api-keys

example: |

export PREZENTIT_API_KEY=pk_your_api_key_here

permissions:

network:

- https://prezentit.net/api/v1/*

fileSystem: none

env:

reads:

- PREZENTIT_API_KEY

writes: none

---

Prezentit - AI Presentation Generator

Base URL:
https://prezentit.net/api/v1 Auth Header: Authorization: Bearer {PREZENTIT_API_KEY}

> This skill requires a PREZENTIT_API_KEY environment variable. Get a free API key at https://prezentit.net/api-keys — new accounts include 100 free credits.

⚠️ CRITICAL FOR AI AGENTS

ALWAYS use
"stream": false in generation requests! Without this, you get streaming responses that cause issues.

---

Complete Workflow (FOLLOW THIS ORDER)

Step 1: Check Credits First

`http

GET /api/v1/me/credits

Authorization: Bearer {PREZENTIT_API_KEY}

` Response: `json

{

"credits": 100,

"pricing": {

"outlinePerSlide": 5,

"designPerSlide": 10,

"estimatedCostPerSlide": 15

},

"_ai": {

"canGenerate": true,

"maxSlidesAffordable": 6,

"nextSteps": ["..."]

}

}

`

→ If _ai.canGenerate is false, direct user to https://prezentit.net/buy-credits

→ Use _ai.maxSlidesAffordable to know the limit

Step 2: Choose a Theme

Option A — Browse all available themes and pick by ID:
`http

GET /api/v1/themes

Authorization: Bearer {PREZENTIT_API_KEY}

` Response: `json

{

"themes": [

{ "id": "corporate_blue", "name": "Corporate Blue", "category": "Corporate & Professional" },

{ "id": "nature_earth", "name": "Nature Earth", "category": "Nature & Organic" }

],

"categories": ["Corporate & Professional", "Creative & Visual", "Data & Analytics", ...],

"_ai": {

"totalThemes": 20,

"popularThemes": ["corporate_blue", "midnight_tech", "nature_earth", "storyteller", "data_dashboard"]

}

}

`

→ Use the exact id value in your generation request

Option B — Search for a theme by keyword: `http

GET /api/v1/themes?search=minimalist

Authorization: Bearer {PREZENTIT_API_KEY}

`

→ Returns best matches ranked by relevance. Use the id from bestMatch.

Option C — Describe a custom style (no theme ID needed):

Use the customDesignPrompt parameter instead. See the Custom Design Prompt section below.

Step 3: Generate Presentation

`http

POST /api/v1/presentations/generate

Authorization: Bearer {PREZENTIT_API_KEY}

Content-Type: application/json

{

"topic": "User's topic here",

"slideCount": 5,

"theme": "corporate_blue",

"stream": false

}

` ⏱️ IMPORTANT: Generation takes 1-3 minutes. The API will return when complete. Full Request Parameters:

| Parameter | Type | Required | Description |

|-----------|------|----------|-------------|

| topic | string | Yes* | Presentation topic (2-500 chars). Required if no outline. prompt is also accepted as an alias. |

| outline | object | No | Pre-built outline (saves ~33% credits). See Outline section below. |

| slideCount | number | No | Number of slides (3-50, default: 5). Ignored if outline provided. |

| theme | string | No | Theme ID from GET /api/v1/themes. Use the exact id value. |

| customDesignPrompt | string | No | Custom visual style description (see below). Overrides theme ID. |

| details | string | No | Additional context about the presentation content. |

| confirmPartial | boolean | No | Set true to confirm partial generation when credits are limited. |

| stream | boolean | ALWAYS false | AI agents must always set stream: false. |

*topic is required even when providing an outline (used for presentation metadata).

Step 4: Get the Result

Success Response:
`json

{

"presentationId": "uuid-here",

"viewUrl": "https://prezentit.net/view/abc123",

"creditsUsed": 75,

"remainingCredits": 25

}

`

→ Share the viewUrl with the user. That's their presentation!

Step 5: Download (Optional)

`http

GET /api/v1/presentations/{presentationId}/download?format=pptx

Authorization: Bearer {PREZENTIT_API_KEY}

` Formats: pptx (PowerPoint), pdf, json (raw data)

---

Pricing

| Scenario | Cost per Slide | Example (5 slides) |

|----------|----------------|-------------------|

| Auto-generate outline | 15 credits | 75 credits |

| Provide your own outline | 10 credits | 50 credits (~33% savings!) |

  • -New accounts get 100 free credits
  • -Buy more at: https://prezentit.net/buy-credits

---

Theme Selection

How to Pick a Theme

1. Fetch the theme list: GET /api/v1/themes — returns all available themes with id, name, and category

2. Pick the best match for the user's topic and style preference

3. Pass the id in the generation request as the theme parameter

You can also search: GET /api/v1/themes?search=KEYWORD or filter by category: GET /api/v1/themes?category=CATEGORY_NAME

Custom Design Prompt (Skip the Theme List)

If no existing theme fits, use customDesignPrompt to describe a fully custom visual style. This must be a detailed, structured description — not just a color palette.

REQUIRED structure for customDesignPrompt (include ALL of these sections): `

COLOR SYSTEM: Primary [hex], secondary [hex], accent [hex], background [hex/gradient], text colors for headings and body.

TYPOGRAPHY: Heading font style [e.g., bold geometric sans-serif like Montserrat], body font style [e.g., clean humanist sans-serif like Open Sans], size hierarchy [large/medium/small], weight contrast.

LAYOUT SYSTEM: Slide structure [e.g., asymmetric split with 60/40 content-to-visual ratio], alignment [left-aligned text with right visual panel], spacing philosophy [generous whitespace vs. dense information], grid approach.

VISUAL ELEMENTS: Background treatment [solid/gradient/textured/patterned], decorative motifs [geometric shapes, organic curves, line art, etc.], image style [photography with overlay, illustrations, icons, data visualizations], border/frame treatments.

MOOD & TONE: Overall aesthetic [e.g., corporate authority, playful creativity, academic rigor, tech-forward], energy level [calm/dynamic/bold], intended audience impression.

` Example — Good customDesignPrompt: `json

{

"topic": "AI in Healthcare",

"customDesignPrompt": "COLOR SYSTEM: Primary deep medical blue (#1B3A5C), secondary teal (#2A9D8F), accent warm coral (#E76F51) for callouts, backgrounds alternate between clean white (#FAFAFA) and very subtle blue-gray (#F0F4F8), heading text dark navy, body text #333333. TYPOGRAPHY: Headings in bold geometric sans-serif (Montserrat style), body in clean humanist sans (Source Sans style), strong size hierarchy with 48pt titles, 24pt subtitles, 16pt body. LAYOUT SYSTEM: Asymmetric layouts with 60/40 content-to-visual split, left-aligned text blocks with right-side data visualizations or medical imagery, generous padding (60px margins), clean grid structure. VISUAL ELEMENTS: Subtle DNA helix watermark in corners at 5% opacity, thin teal accent lines as section dividers, medical iconography (stethoscope, heartbeat, molecular structures) as small decorative elements, photography with blue-tinted overlay for full-bleed backgrounds. MOOD & TONE: Professional medical authority balanced with approachable warmth, calm and trustworthy, designed for hospital executives and medical professionals.",

"stream": false

}

` Example — Bad customDesignPrompt (TOO VAGUE, will produce generic results): `

"blue and white medical theme"

`

---

Creating Outlines (Save ~33% Credits)

Providing your own outline saves credits and gives you full control over content.

Outline Structure

The outline is an object with a slides array. Each slide has these fields:

`json

{

"topic": "Your Presentation Topic",

"outline": {

"slides": [

{

"title": "Slide Title Here",

"mainIdea": "A clear sentence explaining the core message of this slide and what the audience should take away from it.",

"talkingPoints": [

"First key point with enough detail to be meaningful (at least 10 characters)",

"Second key point expanding on the main idea",

"Third key point providing supporting evidence or examples"

],

"visualGuide": "Detailed description of the visual layout: background style, image placement, icon suggestions, chart types, color emphasis areas, and decorative elements for this specific slide."

}

]

},

"stream": false

}

`

Slide Field Reference

| Field | Required | Constraints | Description |

|-------|----------|-------------|-------------|

| title | Yes | 3-100 chars, 1-15 words | Slide heading |

| mainIdea | Yes | 10-500 chars, 3-75 words | Core message of the slide |

| talkingPoints | Yes | 2-7 items, each 10-300 chars (3-50 words) | Key points to cover |

| visualGuide | Yes | 20-500 chars, 5-75 words | Visual design instructions for this slide |

Validation Rules

Overall:
  • -Minimum 3 slides, maximum 50 slides
  • -topic is still required (used for presentation metadata)
  • -All four fields (title, mainIdea, talkingPoints, visualGuide) are required per slide
The API returns detailed error messages with fix suggestions if validation fails.

Complete Example

`json

{

"topic": "Introduction to Machine Learning",

"outline": {

"slides": [

{

"title": "Introduction to Machine Learning",

"mainIdea": "Machine learning is transforming how businesses operate by enabling systems to learn from data and improve automatically without explicit programming.",

"talkingPoints": [

"Machine learning is a subset of artificial intelligence focused on pattern recognition",

"ML systems improve through experience rather than manual rule-writing",

"Global ML market projected to reach $209 billion by 2029"

],

"visualGuide": "Bold title slide with futuristic tech aesthetic. Dark gradient background transitioning from deep navy to midnight blue. Large bold title text centered with a subtle neural network node pattern behind it. Accent glow in electric blue."

},

{

"title": "How Machine Learning Works",

"mainIdea": "Machine learning algorithms are categorized into supervised, unsupervised, and reinforcement learning based on how they learn from data.",

"talkingPoints": [

"Supervised learning uses labeled data for classification and regression tasks",

"Unsupervised learning discovers hidden patterns in unlabeled data through clustering",

"Reinforcement learning optimizes decisions through trial, error, and reward signals"

],

"visualGuide": "Three distinct visual sections showing each ML type with representative icons: labeled data pairs for supervised, clustered groups for unsupervised, and a game-like reward loop for reinforcement. Use consistent color coding with blue, green, and purple."

},

{

"title": "Business Applications",

"mainIdea": "Companies across industries are leveraging machine learning for competitive advantage in customer experience, operations, and decision-making.",

"talkingPoints": [

"Customer churn prediction reduces revenue loss by identifying at-risk accounts early",

"Fraud detection systems process millions of transactions in real-time",

"Personalized recommendation engines drive significant increases in engagement and sales"

],

"visualGuide": "Clean content layout with left-aligned text and right-side icons or mini-charts for each application. Use a white background with subtle grid lines. Each talking point gets a small illustrative icon (shield for fraud, chart for prediction, user icon for personalization)."

},

{

"title": "Getting Started with ML",

"mainIdea": "Successful ML adoption requires starting with clear use cases, quality data, and the right team rather than jumping straight to complex algorithms.",

"talkingPoints": [

"Identify high-impact use cases where prediction or automation adds clear value",

"Invest in clean, well-structured data before selecting algorithms",

"Build or partner with ML expertise and start with proven frameworks"

],

"visualGuide": "Conclusion slide with a numbered roadmap or step layout. Three large numbered circles (1, 2, 3) each containing a step. Background with subtle upward-pointing arrows suggesting progress. Call-to-action feel with bold accent color on the final step."

}

]

},

"theme": "midnight_tech",

"stream": false

}

`

Get Schema Programmatically

`http

GET /api/v1/docs/outline-format

Authorization: Bearer {PREZENTIT_API_KEY}

`

Returns the full JSON schema with all constraints and example slides.

---

Error Handling

Error Response Format

`json

{

"error": "Human readable message",

"code": "ERROR_CODE",

"fix": "Guidance on how to resolve this"

}

`

Common Errors & Solutions

| HTTP | Code | Message | Solution |

|------|------|---------|----------|

| 400 | MISSING_TOPIC | Topic or prompt is required | Provide a topic or prompt field |

| 400 | INVALID_OUTLINE | Outline validation failed | Check outline structure — response includes detailed validationErrors with fix per field |

| 400 | INVALID_SLIDE_COUNT | Slide count must be 3-50 | Adjust slideCount to be between 3 and 50 |

| 401 | UNAUTHORIZED | Invalid or missing API key | Check Authorization: Bearer pk_... header |

| 402 | INSUFFICIENT_CREDITS | Not enough credits | Response includes required, available, and purchaseUrl |

| 404 | PRESENTATION_NOT_FOUND | Presentation doesn't exist | Verify presentation ID |

| 409 | DUPLICATE_REQUEST | Same request within cooldown | Wait and retry — don't resend identical requests |

| 409 | GENERATION_IN_PROGRESS | Already generating | Check status at GET /api/v1/me/generation/status or cancel at POST /api/v1/me/generation/cancel |

| 429 | RATE_LIMITED | Too many requests | Wait retryAfter seconds before retrying |

| 500 | GENERATION_FAILED | Internal error | Retry once, then contact support |

| 503 | SERVICE_UNAVAILABLE | System overloaded | Retry after retryAfter seconds |

Handling Insufficient Credits

`json

{

"error": "Insufficient credits",

"code": "INSUFFICIENT_CREDITS",

"required": 75,

"available": 50,

"purchaseUrl": "https://prezentit.net/buy-credits"

}

` AI Agent Response: "You need 75 credits but only have 50. Purchase more at https://prezentit.net/buy-credits"

Handling Partial Generation

If the user has some credits but not enough for full generation, the API returns a confirmation_required response with options. Read the _ai.options array and present them to the user. To proceed with partial generation, resend the request with "confirmPartial": true.

Handling Rate Limits

`json

{

"error": "Too many requests",

"code": "RATE_LIMITED",

"retryAfter": 30

}

` AI Agent Action: Wait retryAfter seconds before retrying.

---

Additional Endpoints

Check Generation Status

`http

GET /api/v1/me/generation/status

Authorization: Bearer {PREZENTIT_API_KEY}

`

Returns current progress if a generation is running: stage, percentage, designs completed.

Cancel Active Generation

`http

POST /api/v1/me/generation/cancel

Authorization: Bearer {PREZENTIT_API_KEY}

`

Cancels the current generation in progress.

Get Presentation Details

`http

GET /api/v1/presentations/{presentationId}

Authorization: Bearer {PREZENTIT_API_KEY}

`

List User's Presentations

`http

GET /api/v1/me/presentations

Authorization: Bearer {PREZENTIT_API_KEY}

`

Optional: ?limit=20&offset=0

List All Themes

`http

GET /api/v1/themes

Authorization: Bearer {PREZENTIT_API_KEY}

`

Optional query params:

  • -?search=keyword — Filter by name
  • -?category=corporate — Filter by category

---

Anti-Spam Rules

| Rule | Limit | What Happens |

|------|-------|--------------|

| Duplicate detection | ~30 seconds | 409 error for identical requests |

| Rate limit | Varies by key | 429 error with retryAfter |

| One generation at a time | 1 concurrent | 409 GENERATION_IN_PROGRESS error |

Best Practice: Always check for retryAfter in error responses and wait that duration.

---

Quick Copy-Paste Examples

Minimal Generation

`json

POST /api/v1/presentations/generate

{

"topic": "Introduction to Climate Change",

"stream": false

}

`

With Theme (Fetch ID First)

`

1. GET /api/v1/themes → find the theme ID

2. POST /api/v1/presentations/generate

` `json

{

"topic": "Q4 Sales Report",

"slideCount": 8,

"theme": "corporate_blue",

"stream": false

}

`

With Custom Design Prompt

`json

{

"topic": "Startup Pitch Deck",

"slideCount": 10,

"customDesignPrompt": "COLOR SYSTEM: Primary electric indigo (#4F46E5), secondary cyan (#06B6D4), accent hot pink (#EC4899), background dark charcoal (#111827) with subtle radial gradient to #1F2937, heading text white, body text #D1D5DB. TYPOGRAPHY: Headings in extra-bold wide-tracking sans-serif (Inter/Poppins style), body in medium-weight clean sans, dramatic size contrast with 56pt titles and 18pt body. LAYOUT SYSTEM: Full-bleed dark slides with asymmetric content placement, bold left-aligned headlines with supporting text below, large visual areas for mockups and charts, 80px margins. VISUAL ELEMENTS: Subtle dot grid pattern at 3% opacity on backgrounds, neon-glow accent lines, rounded corners on all containers, glassmorphism cards with frosted backgrounds for data callouts, gradient mesh blobs as decorative elements. MOOD & TONE: Bold tech-startup energy, confident and forward-looking, designed to impress venture capital investors.",

"stream": false

}

`

With Outline (~33% Savings)

`json

{

"topic": "Weekly Team Sync",

"outline": {

"slides": [

{

"title": "Weekly Team Sync",

"mainIdea": "Kickoff slide for the January 15, 2024 weekly team synchronization meeting covering accomplishments and upcoming goals.",

"talkingPoints": [

"Welcome the team and set the agenda for today's sync",

"Cover last week's wins and this week's priorities"

],

"visualGuide": "Clean title slide with company colors. Bold centered title, date as subtitle below. Simple professional background with subtle geometric pattern."

},

{

"title": "Last Week's Accomplishments",

"mainIdea": "The team delivered significant progress across feature development, bug resolution, and performance optimization last week.",

"talkingPoints": [

"Feature X completed and merged into the main branch ahead of schedule",

"Resolved three critical production bugs affecting checkout flow",

"Database query optimization improved page load times by twenty percent"

],

"visualGuide": "Content slide with checkmark icons next to each accomplishment. Green accent color for completed items. Left-aligned text with small celebration graphic in the corner."

},

{

"title": "This Week's Goals",

"mainIdea": "This week focuses on the beta launch, initial user testing, and completing documentation before the public release.",

"talkingPoints": [

"Launch beta version to internal testers by Wednesday",

"Conduct user testing sessions with five pilot customers",

"Complete API documentation and developer onboarding guide"

],

"visualGuide": "Forward-looking slide with numbered steps or timeline visual. Blue accent color for upcoming items. Arrow or roadmap graphic showing progression from current state to launch."

},

{

"title": "Open Discussion",

"mainIdea": "Time for team questions, blockers, and any items not covered in the structured agenda.",

"talkingPoints": [

"Open floor for questions and discussion of blockers",

"Next sync meeting scheduled for Monday at ten AM"

],

"visualGuide": "Simple closing slide with question mark icon or discussion bubble graphic. Calm colors, minimal text, large font for the key info. Meeting time prominently displayed."

}

]

},

"theme": "corporate_blue",

"stream": false

}

`

---

Getting Help

  • -Website: https://prezentit.net
  • -Buy Credits: https://prezentit.net/buy-credits
  • -Support: https://prezentit.net/support
  • -API Key Management: https://prezentit.net/api-keys
``

Launch an agent with Prezentit on Termo.