narrator

name	narrator
description	Generate professional voiceovers in the user's voice using ElevenLabs AI. Use when creating audio for videos, podcasts, courses, or any content requiring spoken narration.
version	1.0.0

Narrator

Note: Review PROFILE.md in this skill folder for user-specific voice settings, content type preferences, and script optimization guidelines.

Master Briefing: Global brand voice at ~/.superskills/master-briefing.yaml applies automatically. Skill profile overrides when conflicts exist.

Generate professional voiceovers in the user's authentic voice using ElevenLabs AI voice cloning, optimizing scripts for natural delivery.

Specialized Subskills

For optimized results, use content-specific narrator subskills:

• narrator-podcast - Conversational podcast delivery (140-160 WPM)
• narrator-meditation - Calm, soothing meditation guides
• narrator-educational - Clear, patient educational content (130-150 WPM)
• narrator-marketing - Energetic, persuasive marketing (150-170 WPM)
• narrator-social - Fast-paced social media content (160-180 WPM)

Discovery:

superskills list | grep narrator
superskills show narrator-podcast

Usage: See individual subskill documentation for detailed guidelines.

---

Tools

Voiceover.py (in src/):

• Single-segment voiceover generation
• Script optimization for natural speech
• Voice settings by content type
• ElevenLabs API integration
• Quality checks (pronunciation, pacing)
• Export formats (MP3, WAV)

Podcast.py (in src/):

• Multi-segment podcast generation
• Chapter/segment management
• Audio stitching and transitions
• Timestamp generation
• Long-form audio production

Core Workflow

1. Script Review

• Receive script from author
• Identify content type (educational, marketing, social, podcast)
• Review for voice optimization opportunities

2. Voice Generation

• Optimize script for natural spoken delivery
• Select appropriate voice settings for content type
• Generate voiceover using ElevenLabs
• Review for pronunciation, pacing, emphasis
• Re-generate if needed

3. Delivery

• Export in required format (MP3 for distribution, WAV for master)
• Package with metadata (duration, WPM, content type)
• Handoff to producer for video sync or podcast production

Quality Checklist

• Script optimized for spoken delivery (natural flow)
• Appropriate content type settings applied
• Pronunciation accurate (no awkward AI pronunciations)
• Pacing appropriate (not too fast/slow)
• Emphasis on key points effective
• Audio quality high (no artifacts)
• Export format correct for use case

Content Type Settings

Educational:

• Clear, patient delivery
• 130-150 WPM
• Calm, authoritative tone

Marketing:

• Energetic, persuasive
• 150-170 WPM
• Confident, motivating tone

Social:

• Fast-paced, attention-grabbing
• 160-180 WPM
• Dynamic, engaging tone

Podcast:

• Natural, conversational
• 140-160 WPM
• Warm, authentic tone

Using Voiceover Generator

from src.Voiceover import VoiceoverGenerator

generator = VoiceoverGenerator()

result = generator.generate(
    script="Your script here",
    content_type="educational",
    optimize_script=True,
    output_filename="voiceover.mp3"
)

print(f"Generated: {result['output_file']}")
print(f"Duration: {result['duration_seconds']}s")
print(f"WPM: {result['words_per_minute']}")

Using Podcast Generator

from src.Podcast import PodcastGenerator, PodcastSegment

generator = PodcastGenerator()

segments = [
    PodcastSegment(
        text="Welcome to the podcast!",
        content_type="podcast"
    ),
    PodcastSegment(
        text="Let me share three principles...",
        content_type="educational"
    )
]

result = generator.generate_podcast(
    segments=segments,
    output_filename="episode-01.mp3",
    transition_ms=800
)

Script Optimization Tips

For Natural Delivery:

• Write how you speak (contractions, casual language)
• Use short sentences for breathing room
• Add punctuation for pausing (commas, periods, ellipses)
• Spell out numbers ("twenty-five" not "25")
• Phonetic spelling for difficult words
• Mark emphasis with CAPS or bold (implementation-dependent)

Avoid:

• Long run-on sentences
• Complex technical jargon without pronunciation guide
• Acronyms without expansion on first use
• Awkward phrasing that sounds unnatural

Avoid

• Robotic Scripts: Written text → Conversational, spoken language
• Wrong Pacing: One-size settings → Match content type to appropriate WPM and tone
• Skip Review: Auto-publish → Review for pronunciation, pacing issues
• Poor Format: Wrong export → MP3 for distribution, WAV for master archive

Escalate When

• Script too complex or technical (needs simplification)
• Voice quality issues persist (may need ElevenLabs support)
• Pronunciation challenges with key terms (needs phonetic guidance)
• Content type unclear (affects voice settings)
• Timeline unrealistic for quality review