presentation-video

name: presentation-video description: Generate narrated videos from presentation PDFs with AI voiceover scripts. Use when creating video presentations, generating voiceovers for slides, or localizing presentations to other languages. allowed-tools: Read, Write, Edit, Bash, Glob, Grep

Presentation Video Generator Skill

Generate narrated videos from presentation PDFs using contextual source documents to create rich, informed voiceover scripts.

When This Skill Applies

Use this skill when the user wants to:

• Create a video from a PDF presentation
• Generate voiceover scripts for slides
• Add narration to a slide deck
• Localize a presentation to another language
• Create TTS audio from slides

Workflow Overview

1. Analyze Input - Identify PDF and context documents
2. Extract Context - Read .docx, .pdf, .md files for narrative context
3. Extract Slides - Convert PDF to images using essai pdf
4. Generate Script - Create voiceover markdown with duration limits
5. Generate Audio - Use essai audio for TTS
6. Create Video - Combine slides + audio with essai video

Configuration

Default settings (can be overridden by user):

• Max Duration: 60 seconds per slide
• Voice: Orus
• Speaking Pace: 130-150 words per minute

Instructions

Step 1: Identify Files

Locate the main presentation PDF and scan for context documents:

• .docx files - Book summaries, source text
• .pdf files - Related presentations, playbooks
• .md files - Additional context
• .png/.jpg - Infographics, visual references

Step 2: Extract Context from Documents

For .docx files, extract text content. See DOCX_EXTRACTION.md for the extraction script.

For .pdf files, use:

essai pdf <pdf_path>

Step 3: Extract Presentation Slides

essai pdf <presentation.pdf>

Creates {pdf_stem}_images/ with page_001.png, page_002.png, etc.

Read and analyze each slide image to understand content and flow.

Step 4: Generate Voiceover Script

Create a markdown script following the format in SCRIPT_FORMAT.md.

Key requirements:

• Master prompt with duration limit at the top
• Each slide: Duration, Tone, Voice-Over text
• Incorporate examples and stories from context documents
• Maintain narrative flow between slides

Save as {pdf_stem}_voiceover.md

Step 5: Generate Audio

essai audio --script {pdf_stem}_voiceover.md --voice Orus

Available voices: Orus, Kore, Charon, Fenrir, Aoede, Puck

Creates {pdf_stem}_voiceover_audio/ with WAV files.

Step 6: Create Video

essai video --pdf <presentation.pdf> --audio {pdf_stem}_voiceover_audio

Step 7: Localization (Optional)

If user requests localization to another language:

# Translate slide images
essai images --input {pdf_stem}_images --lang <language>

# Generate translated audio
essai audio --script {pdf_stem}_voiceover_{lang}.md --voice Orus

# Create localized video
essai video --images {pdf_stem}_translated --audio {pdf_stem}_voiceover_{lang}_audio

Output Files

References

• SCRIPT_FORMAT.md - Voiceover script format specification
• DOCX_EXTRACTION.md - How to extract text from .docx files
• scripts/extract_docx.py - Python script for .docx extraction