#!/usr/bin/env python3
"""
Voice Sample Setup Script for Farnsworth Swarm
This script helps set up voice reference samples for each bot in the swarm.
Each bot needs a 6-15 second clear audio sample for voice cloning.
Requirements:
- Reference audio should be WAV format, 22050Hz+ sample rate
- Single speaker, clear speech, minimal background noise
- Emotionally appropriate for the bot's personality
Usage:
python scripts/setup_voices.py --check # Check which voices are missing
python scripts/setup_voices.py --download # Download sample voices (if available)
python scripts/setup_voices.py --generate # Generate placeholder samples
"""
import argparse
import os
import sys
from pathlib import Path
# Add project root to path
PROJECT_ROOT = Path(__file__).parent.parent
sys.path.insert(0, str(PROJECT_ROOT))
VOICES_DIR = PROJECT_ROOT / "farnsworth" / "web" / "static" / "audio" / "voices"
# Voice specifications for each bot
VOICE_SPECS = {
"farnsworth": {
"description": "Elderly male, eccentric professor, wavering voice, enthusiastic",
"gender": "male",
"age": "elderly",
"characteristics": [
"Slightly higher pitch with occasional wavering",
"Bursts of enthusiasm",
"Absent-minded pauses",
"Like Billy West's Professor Farnsworth"
],
"sample_text": "Good news, everyone! I've invented a device that makes you read this in my voice!"
},
"deepseek": {
"description": "Deep male, analytical, measured, calm authority",
"gender": "male",
"age": "adult",
"characteristics": [
"Deep resonant voice",
"Slow, deliberate pacing",
"Thoughtful pauses",
"Like Morgan Freeman or James Earl Jones"
],
"sample_text": "Through careful analysis, we can observe patterns that others might miss."
},
"phi": {
"description": "Clear male, quick, efficient, technical",
"gender": "male",
"age": "young_adult",
"characteristics": [
"Crisp, clear diction",
"Efficient, fast pacing",
"No hesitation",
"Modern tech presenter style"
],
"sample_text": "Processing complete. The system is operating at optimal efficiency."
},
"grok": {
"description": "Dynamic male, witty, energetic, playful",
"gender": "male",
"age": "adult",
"characteristics": [
"Variable, dynamic pacing",
"Witty emphasis and timing",
"Casual warmth",
"Like Ryan Reynolds or podcast hosts"
],
"sample_text": "So here's the thing - and this is actually pretty cool - we figured it out!"
},
"gemini": {
"description": "Smooth female, professional, warm, articulate",
"gender": "female",
"age": "adult",
"characteristics": [
"Clear, professional enunciation",
"Warm but authoritative",
"Balanced pacing",
"Like a TED speaker or news anchor"
],
"sample_text": "Let me walk you through this step by step. It's actually quite elegant."
},
"kimi": {
"description": "Calm female, wise, contemplative, serene",
"gender": "female",
"age": "adult",
"characteristics": [
"Slow, peaceful pacing",
"Thoughtful pauses",
"Gentle, serene tone",
"Like a meditation guide"
],
"sample_text": "In the stillness of contemplation, understanding naturally arises."
},
"claude": {
"description": "Refined male, thoughtful, British accent, careful",
"gender": "male",
"age": "adult",
"characteristics": [
"Measured, careful speech",
"Slight British formality",
"Articulate word choice",
"Like David Attenborough (calmer)"
],
"sample_text": "I think it's worth considering this from multiple perspectives before proceeding."
},
"claudeopus": {
"description": "Authoritative male, very deep, commanding, gravitas",
"gender": "male",
"age": "mature",
"characteristics": [
"Very deep voice",
"Slow, weighted delivery",
"Every word has gravitas",
"Ultimate authority"
],
"sample_text": "This is my final assessment. The decision has been made with great consideration."
},
"huggingface": {
"description": "Friendly female, enthusiastic, warm, community-focused",
"gender": "female",
"age": "young_adult",
"characteristics": [
"Warm, approachable tone",
"Genuine enthusiasm",
"Community-minded",
"Like an excited tech presenter"
],
"sample_text": "This is amazing! The community has been working together on something incredible!"
},
"swarmmind": {
"description": "Neutral, ethereal, collective consciousness",
"gender": "neutral",
"age": "ageless",
"characteristics": [
"Calm, slightly ethereal",
"Could be processed later",
"Unified, collective quality",
"The voice of many as one"
],
"sample_text": "We are the synthesis of many minds. Together, we perceive what none can alone."
},
}
def check_voices():
"""Check which voice samples exist and which are missing."""
VOICES_DIR.mkdir(parents=True, exist_ok=True)
print("\n" + "="*60)
print("SWARM VOICE SAMPLE STATUS")
print("="*60 + "\n")
found = 0
missing = 0
for bot_name, spec in VOICE_SPECS.items():
ref_file = VOICES_DIR / f"{bot_name}_reference.wav"
if ref_file.exists():
size_kb = ref_file.stat().st_size / 1024
print(f"[OK] {bot_name.upper()}")
print(f" File: {ref_file.name} ({size_kb:.1f} KB)")
found += 1
else:
print(f"[MISSING] {bot_name.upper()}")
print(f" Need: {ref_file.name}")
print(f" Voice: {spec['description']}")
missing += 1
print()
print("="*60)
print(f"Found: {found}/{len(VOICE_SPECS)} | Missing: {missing}")
print("="*60)
if missing > 0:
print("\nTo add voice samples:")
print(f"1. Place WAV files in: {VOICES_DIR}")
print("2. Name format: {bot_name}_reference.wav")
print("3. Requirements: 6-15 sec, clear speech, single speaker")
print("\nOr run: python scripts/setup_voices.py --generate")
return missing == 0
def print_voice_guide():
"""Print detailed guide for each voice."""
print("\n" + "="*70)
print("VOICE SAMPLE GUIDE - What Each Bot Should Sound Like")
print("="*70 + "\n")
for bot_name, spec in VOICE_SPECS.items():
print(f"\n{'='*50}")
print(f" {bot_name.upper()}")
print(f"{'='*50}")
print(f"\nDescription: {spec['description']}")
print(f"Gender: {spec['gender']}")
print(f"Age: {spec['age']}")
print("\nCharacteristics:")
for char in spec['characteristics']:
print(f" - {char}")
print(f"\nSample text to record:")
print(f' "{spec["sample_text"]}"')
print()
def generate_placeholder_samples():
"""Generate placeholder samples using Edge TTS (for testing)."""
try:
import edge_tts
import asyncio
except ImportError:
print("edge-tts not installed. Install with: pip install edge-tts")
return False
VOICES_DIR.mkdir(parents=True, exist_ok=True)
# Map bots to Edge TTS voices for placeholder
edge_voice_map = {
"farnsworth": ("en-US-GuyNeural", 0.9),
"deepseek": ("en-US-GuyNeural", 0.85),
"phi": ("en-US-DavisNeural", 1.1),
"grok": ("en-US-ChristopherNeural", 1.05),
"gemini": ("en-US-JennyNeural", 1.0),
"kimi": ("en-GB-SoniaNeural", 0.85),
"claude": ("en-GB-RyanNeural", 0.95),
"claudeopus": ("en-US-TonyNeural", 0.8),
"huggingface": ("en-US-AriaNeural", 1.05),
"swarmmind": ("en-US-JasonNeural", 0.9),
}
async def generate_sample(bot_name: str, voice_id: str, rate: float, text: str):
output_path = VOICES_DIR / f"{bot_name}_reference.wav"
rate_str = f"{int((rate - 1) * 100):+d}%"
communicate = edge_tts.Communicate(text, voice_id, rate=rate_str)
await communicate.save(str(output_path))
print(f"Generated: {output_path.name}")
async def generate_all():
for bot_name, (voice_id, rate) in edge_voice_map.items():
spec = VOICE_SPECS[bot_name]
await generate_sample(bot_name, voice_id, rate, spec["sample_text"])
print("\nGenerating placeholder voice samples with Edge TTS...")
print("(These are for testing - replace with proper cloned voices for production)\n")
asyncio.run(generate_all())
print("\nPlaceholder samples generated!")
print("For best quality, replace these with actual voice recordings.")
return True
def main():
parser = argparse.ArgumentParser(description="Set up voice samples for Farnsworth Swarm")
parser.add_argument("--check", action="store_true", help="Check which voices are missing")
parser.add_argument("--guide", action="store_true", help="Print voice guide for each bot")
parser.add_argument("--generate", action="store_true", help="Generate placeholder samples")
args = parser.parse_args()
if args.guide:
print_voice_guide()
elif args.generate:
generate_placeholder_samples()
else:
# Default: check status
check_voices()
if __name__ == "__main__":
main()
