Slide Generation

Slide mode can automatically generate complete presentation slides using AI, with style analysis and page-by-page generation.

Interface Overview

Slides mode offers two main approaches: plan the content yourself, or let AI plan it for you.

Scenario 1: Plan Content Yourself

When you already know what each page should contain, you can enter content directly in the "Presentation Content" area, using --- (three dashes) to separate different pages.

Example: Creating a FIDO Passkey Introduction

Let's create a 3-page presentation to introduce FIDO Passkey:

Cover: What is FIDO Passkey?
Subtitle: A more secure and convenient passwordless login method

---

How Passkey Works
- Uses public key cryptography
- Private key stored securely on device
- Verified through biometrics (fingerprint, Face ID)
- No need to remember complex passwords

---

Advantages of Passkey
- Prevents phishing: private key never leaves the device
- Cross-device sync: syncs via iCloud/Google account
- Better user experience: one-tap login, no password needed

The system will automatically split this into three pages based on ---, showing "3 pages" at the bottom.

Steps

1. Select "Slides" mode
2. Enter content for each page in "Presentation Content", separated by --- (max 30 pages)
3. (Optional) Enter "Global Description" - background context about your presentation that helps AI understand the overall theme (this text won't appear on slides)
4. Set quality (1K, 2K, 4K) and aspect ratio (16:9, 4:3, 1:1)
5. Choose design style (AI Analysis or Manual Input)
6. Click "Generate"

Global Description

Use Global Description to provide context like "This is an internal tech sharing session for engineers" or "Target audience is marketing team, prefer business style". This helps AI generate more appropriate visuals without cluttering individual slide content.

Scenario 2: Let AI Plan for You

When you have a long document (like official documentation, technical specs, meeting notes) to convert into slides, you can use the "AI Planning" feature to let AI analyze the content and plan key points for each page.

About API Key Usage

AI Planning is a text processing feature that prioritizes using your Free Tier API Key (if configured). It only switches to your paid key when the Free Tier quota is exhausted. See API Key Management for details.

Example: Converting FIDO Official Documentation to Slides

1. Click the "AI Planning" button next to "Presentation Content"
2. Paste the full document content in the popup (e.g., FIDO Alliance technical whitepaper)
3. Click "Start Planning"
4. AI will analyze and split the content into multiple slide pages

Tip

AI Planning works best for longer documents. If you only have brief bullet points, Scenario 1 (planning yourself) is more efficient.

Generation Flow

Slides mode uses a sequential generation strategy:

1. Generate first page (cover) first
2. Use first page as reference to maintain style consistency
3. Generate subsequent pages in order

This ensures visual style consistency throughout the presentation.

Page Editing & Selective Regeneration

After generation, if you only need to adjust a few pages, there's no need to regenerate everything.

Inline Editing on Page Cards

Each page card has an edit button (✏️). Click it to directly modify that page's content. Changes automatically sync back to the "Presentation Content" textarea above.

Likewise, if you edit content in the textarea above, the corresponding page card below will also update.

Change Detection

The system automatically tracks three types of changes per page:

Change Type	Trigger
Content	Modified page text content
Style	Modified page style guide or global style
Narration	Modified narration script

Modified pages display a yellow border and a "Modified" badge. If you revert content back to what it was when generated, the modification marker automatically disappears.

Selective Regeneration

When pages are marked as modified, the generate button area splits into two buttons:

1. Generate Modified Pages (primary): Only regenerate modified pages, showing a detailed breakdown (e.g., "slides p.2, 5, audio p.3")
2. Regenerate All (secondary): Regenerate all pages from scratch

Save API Quota

If you only modified 1 page's text, "Generate Modified Pages" will only use 1 API call instead of 15.

Viewing Page Images

Click a page card's thumbnail to open the lightbox for a full-size view, with narration audio playback and transcript panel support.

Design Style

In the "Design Style" section, you can choose from two methods to set your presentation's visual style:

AI Analysis

Click "AI Analysis", then you can:

1. Select analysis model: Gemini 3 Flash (faster) or Gemini 3.1 Pro (more accurate)
2. Enter style preferences (optional): Describe what you want or don't want, e.g.:
   • ✓ Want: Modern minimalist, blue color scheme
   • ✗ Don't want: Gradient backgrounds, excessive decorations
3. Click "Analyze & Plan Style"

AI will analyze your presentation content and style preferences to generate suitable design style suggestions:

• Color scheme: Primary, secondary, background colors
• Layout: Title position, content blocks, margins
• Typography: Title size, body text size
• Visual elements: Chart style, icon style

You can edit and adjust the AI's suggestions after they're generated.

About API Key Usage

Style analysis is a text processing feature that prioritizes using your Free Tier API Key. See API Key Management for details.

Manual Input

If you already have a clear style in mind, choose "Manual Input" and directly describe your desired design style:

Modern minimalist style with navy blue and white background,
sans-serif fonts, clean and organized layout

Global Style vs Page Style

Presentation styles work at two levels:

1. Global Style: Base design style applied to all pages
2. Page Style Guide: Each page can have additional style adjustments

For example, you can set the global style to "professional business" but specify "use blue-green data visualization" for chart pages.

Tip

If a page has special content (charts, quotes, timelines), you can specify custom styling in that page's "Page Style Guide" to help AI handle it better.

Global Reference Images

You can upload up to 5 reference images that will be applied to all page generations, helping AI better understand your desired visual style or brand elements.

Prompt Structure (Technical Details)

The system combines your inputs into a structured prompt sent to AI for image generation. Here's the actual prompt structure for each page:

# Slide Generation Task

Generate a presentation slide image for **Page {number} of {total}**.

## PRESENTATION OVERVIEW
{Global Description}

## DESIGN STYLE GUIDE

### Global Style
{Global Style}

### Page-Specific Adjustments
{Page Style Guide}

## SLIDE CONTENT
{Page Content}

## DESIGN REQUIREMENTS
(System-added design guidelines)

## STRICT CONSTRAINTS
(System-added constraints)

This structure ensures:

• Global Description provides background context (not displayed on slides)
• Global Style maintains visual consistency across all slides
• Page Style Guide allows flexibility for individual pages
• Page Content is the actual text rendered on the slide

Page Types

Since this is an image generation model, the types of pages you can create are only limited by your imagination. Here are some common examples:

Type	Description
Cover	Title, subtitle, date
Table of Contents	Presentation outline
Content Page	Title + bullet points
Chart Page	Data visualization
Comparison Page	Two or multi-column comparison
Timeline	History, milestones
Quote Page	Famous quotes, key takeaways
Ending Page	Summary, Q&A, contact info

As long as you can clearly describe the layout in your prompt, the model can generate it for you.

Prompt Tips

Specify Content Clearly

Page 3: Quarterly Performance
- Q1: Revenue $1.2M, growth 15%
- Q2: Revenue $1.35M, growth 12%
- Q3: Revenue $1.28M, decline 5%
- Q4: Revenue $1.5M, growth 17%
Use bar chart visualization

Specify Style Preferences

Overall style: Professional business
Colors: Navy blue primary, white background
Avoid: Excessive decoration, cartoon elements

Narration Audio

Slides mode supports generating narration audio (Text-to-Speech) for each page, ideal for creating narrated presentations or pre-recorded demos.

Enabling Narration

1. After completing style analysis and confirming, the "Narration" section will appear
2. Toggle "Enable Narration" on
3. Configure language, speaker mode, conversation style, etc.

Speaker Modes

Mode	Description
Single	One speaker monologue
Dual	Two speakers in conversation (discussion, critical, or debate style)

Generation Flow

1. Generate Scripts: AI automatically writes narration scripts for each page based on slide content
2. Review/Edit: Expand each page's script to fine-tune
3. Generate Audio: When you click "Generate", image generation and TTS audio generation run in parallel

Narrative Structure

AI automatically arranges a narrative arc: the first page includes a cohesive introduction (greeting, topic overview), and the last page includes a closing statement (summary, takeaway). For single-page presentations, both opening and closing are included on the same page.

Playback & Download

• Live Preview: A mini audio player appears below each image card in the generation results
• Lightbox Playback: Click an image to open Lightbox, with an audio player at the bottom
• Transcript Panel: Click the transcript button in the Lightbox toolbar (or press T) to open a floating transcript panel that auto-updates as you navigate pages. In dual-speaker mode, speakers are color-coded. The panel supports drag to reposition and top-right corner drag to resize; position and size are remembered automatically
• Download Options:
  • ZIP download automatically includes audio files (narration-1.mp3, narration-2.mp3…)
  • Lightbox download menu has a "Narration Audio" section for downloading current page or all audio
  • MP4 download merges all page images and narration audio into a single video (with quality selection; requires WebCodecs support; not available in Firefox)
  • PDF download contains images only (no audio)

Audio Format

Audio is stored in MP3 format by default (64kbps). If MP3 encoding fails, it automatically falls back to WAV format without affecting playback.

About API Key Usage

Script generation and TTS audio generation are text processing features that prioritize using your Free Tier API Key. See API Key Management for details.

Generated Results

After generation, each image is displayed in the preview area. If narration is enabled, a mini audio player appears below each image.

Export Options

After generation, you can:

• Download as ZIP (all page images + narration audio)
• Download as PDF (images only)
• Download as MP4 video (images + narration audio merged into a video, using WebCodecs H.264/AAC encoding)

MP4 Export

MP4 export requires browser support for the WebCodecs API. Currently supported in Chrome and Edge; not available in Firefox (the button is automatically hidden).

When you click the MP4 button, a settings dialog appears:

• Quality: Low (4 Mbps), Medium (8 Mbps, default), High (12 Mbps)
• Resolution: 720p / 1080p / 1440p / 4K available depending on source image dimensions
• Narration Speed: 1x–4x adjustment with pitch preservation. Includes a slider, number input, and preset buttons (1 / 1.25 / 1.5 / 1.75 / 2 / 3x)

All settings are remembered for next time. Pages without narration are filled with 5 seconds of silence; pages with narration last as long as the (speed-adjusted) audio.

Need Editable PPTX?

If you need to convert your slides to editable PPTX format, use the Slide Conversion Tool.

Difference from Slide Conversion

Feature	Slide Generation	Slide Conversion
Input	Text description	PDF file
Purpose	Create slides from scratch	Convert existing PDF to editable format
AI Role	Generate content and design	OCR recognition and background removal

Next Steps

• Slide Conversion - Convert PDF to PPTX
• Diagram Generation - Generate charts for slides
• Image Generation - Generate illustration assets