audio-script-writer

Files (5)

• SKILL.md
• _meta.json
• guidelines.md · references/
• requirements.txt
• main.py · scripts/

SKILL.md (first lines)

# Audio Script Writer

## Overview

Content transformation tool that converts written medical and scientific materials into professionally structured audio scripts suitable for podcasts, educational videos, audiobooks, and voiceover narration.

**Key Capabilities:**
- **Format Conversion**: Research papers → podcast scripts
- **Spoken Word Optimization**: Sentence restructuring for listening
- **Pronunciation Guides**: Medical terminology phonetic spelling
- **Timing Estimation**: Duration calculations for production planning
- **Multi-Format Output**: Podcast, video, lecture, audiobook templates
- **Voice Direction**: Tone, pace, and emphasis cues for narrators

## When to Use

**✅ Use this skill when:**
- Creating medical education podcasts from journal articles
- Converting conference presentations to video scripts
- Developing audiobook versions of medical textbooks
- Scripting patient education audio materials
- Producing research summary videos for social media
- Adapting written case reports for audio case studies
- Creating voiceover scripts for e-learning modules

**❌ Do NOT use when:**
- Live presentation without script → Use improvisation
- Highly visual content (surgery videos) → Use visual-focused tools
- Interactive audio (Q&A format) → Use dialogue scripting tools
- Music or sound design planning → Use audio production software
- Voice recording itself → This creates scripts, not audio

**Integration:**
- **Upstream**: `abstract-summarizer` (content condensation), `lay-summary-gen` (patient-friendly language)
- **Downstream**: `medical-translation` (multi-language scripts), `voice-cloning-tool` (AI narration)

## Core Capabilities

### 1. Spoken Word Transformation

Convert written text to conversational audio style:

```python
from scripts.audio_writer import AudioScriptWriter

writer = AudioScriptWriter()

# Transform written content
script = writer.convert_to_audio(
    source_text=research_paper,
    format="podcast",  # podcast, video, lecture, audiobook
    target_audience="medical_students",
    duration_minutes=15
)

print(script.spoken_text)
# Converts: "The pathophysiology of diabetes mellitus involves..."
# To: "So what exactly happens in diabetes? Well, it all starts when..."
```

**Transformation Rules:**
| Written Style | Audio Style | Example |
|---------------|-------------|---------|
| "Furthermore" | "Plus" | Less formal transitions |
| " et al." | "and their colleagues" | Expand abbreviations |
| Numbers in text | Spoken numbers | "15%" → "15 percent" |
| Long sentences | 15-20 word max | Break into digestible chunks |
| Passive voice | Active voice | "was observed" → "we saw" |
| Citations | Omit or footnote | "(Smith et al., 2024)" → [reference tone] |

### 2. Pronunciation Guide Generation

Create phonetic spelling for medical terms:

```python
# Generate pronunciation guide
pronunciation = writer.create_pronunciation_guide(
    text=script,
    include_phonetic=True,
    include_syllables=True
)

# Output:
# "Hyperlipidemia: hi-per-lip-i-DEE-mee-uh"
# "Metformin: met-FOR-min"
# "Atherosclerosis: ath-er-oh-skleh-ROH-sis"
```

**Guide Elements:**
- **Phonetic Spelling**: IPA or simplified phonetics
- **Syllable Breaks**: hy-per-ten-sion
- **Emphasis Marking**: Primary stress (CAPS), secondary stress
- **Alternative Pronunciations**: Regional variations (UK vs US)
- **Sound-Alikes**: "rhymes with..." for difficult terms

### 3. Timing and Pacing

Calculate speaking duration and mark pacing cues:

```python
# Analyze timing
timing = writer.calculate_timing(
    script=script,
    speaking_rate="conversational",  # slow, conversational, fast
    include_pauses=True
)

print(f"Estimated duration: {timing.duration_minutes} minutes")
print(f"Word count: {timing.word_count}")
print(f"Pace: {timing.words_per_minute} WPM")
```

**Speaking Rates:**
| Style | WPM | Use Case |
|-------|-----|----------|
| **Slow/Educational** | 120-130 | Patient education, complex topics |
| **Conversational** | 140-160 | Podcasts, general audience |
| **Fast/News** | 170-190 | Time-constrained content |
| **Variable** | Varies | Dynamic pacing with pauses |

From LeoYeAI/openclaw-master-skills, MIT.

More from LeoYeAI/openclaw-master-skills