# Ivx Cf Voice Generation

> Design and implement text-to-speech and voice synthesis pipelines. Use when generating voiceovers, cloning voices, or designing audio content.

- Skill: `intelli-verse-x/ivx-cf-voice-generation` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add intelli-verse-x/ivx-cf-voice-generation`
- Raw SKILL.md: https://api.skillmd.com/api/skills/intelli-verse-x/ivx-cf-voice-generation/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: intelli-verse-x (https://skillmd.com/u/intelli-verse-x)
- Updated: 2026-09-08
- Page: https://skillmd.com/skills/intelli-verse-x/ivx-cf-voice-generation

---

# Voice Generation Skill

## Purpose

Generate high-quality voice audio for video and interactive content.

## Pipeline

1. **Script**: Finalized text with timing marks
2. **Voice Selection**: Match character/brand voice
3. **Generation**: TTS with emotion/style control
4. **Post-Processing**: Normalize, EQ, de-ess
5. **Sync**: Align with video timing

## Tools

| Tool | Use Case | Quality |
|------|----------|---------|
| ElevenLabs | Professional voiceovers | High |
| Kokoro | Fast, lightweight TTS | Medium |
| F5-TTS | Voice cloning | High |
| Chatterbox | Conversational | Medium |

## CF Integration

```python
from utils.pipeline.voice_consistency import assign_voice
from utils.pipeline.voice_generator import generate_speech

voice_id = assign_voice(
    character_id=character_id,
    brand_id=brand_id,
    emotion="neutral",
)

audio = generate_speech(
    text=script,
    voice_id=voice_id,
    speed=1.0,
    stability=0.5,
)
```

## Quality Gates

- [ ] Pronunciation correct
- [ ] Emotion matches scene
- [ ] Consistent with previous episodes
- [ ] No audio artifacts (clipping, noise)
- [ ] Synced to video (if applicable)

## Cost

- ElevenLabs: ~$0.30/min (high quality)
- Self-hosted Kokoro: ~$0.01/min (GPU cost)
- Batch generation for cost savings