Voice cloning — clones the principal's voice via ElevenLabs API from 3 MP3 samples
Voice cloning — Instant Voice Clone from 3 MP3 samples (30 seconds each)
Text to Speech — converts VSL scripts, podcast intros, video narrations to MP3
Outbound calls — automated follow-up calls to leads in the principal's voice
Inbound calls — answers 24/7, qualifies prospects, books calls, reports outcomes
Telegram notifications — alerts on hot leads, call summaries, audio ready

Required Credentials

REQUIRED (add to .env before first use):
  ELEVENLABS_API_KEY=sk_...      → elevenlabs.io → Developers → API Keys
  ELEVENLABS_VOICE_ID=...        → created after voice cloning (see SKILL.md)
  TELEGRAM_BOT_TOKEN=...         → agent's existing Telegram bot
  TELEGRAM_CHAT_ID=...           → agent's existing chat ID

OPTIONAL (for phone calls):
  TWILIO_ACCOUNT_SID=...         → console.twilio.com
  TWILIO_AUTH_TOKEN=...          → console.twilio.com
  TWILIO_PHONE_NUMBER=+1...      → buy a number on Twilio (~$1/month)

The voice cloning setup guide (PATH A terminal + PATH B browser) is in SKILL.md — the agent follows it step by step to create ELEVENLABS_API_KEY and ELEVENLABS_VOICE_ID on first run.

What You Provide

Minimum (TTS only):
  → ELEVENLABS_API_KEY in .env (from elevenlabs.io/app/settings/api-keys)
  → 3 MP3 voice samples in /workspace/voice/samples/

For calls (add):
  → TWILIO_ACCOUNT_SID + TWILIO_AUTH_TOKEN + TWILIO_PHONE_NUMBER

What the Agent Retrieves Automatically

ELEVENLABS_API_KEY   → created in elevenlabs.io dashboard
ELEVENLABS_VOICE_ID  → created after uploading your samples
ELEVENLABS_AGENT_ID  → created if Twilio credentials present

Files

File	Role
`SKILL.md`	Full voice system — setup, TTS, calls
`README.md`	This file
`scripts/voice_generator.py`	Python CLI — TTS generation, call management
`templates/agent_prompt.md`	Conversation instructions for the ElevenLabs agent
`templates/config.json`	Config structure template
`references/setup_guide.md`	Browser dashboard guide (optional — for manual setup)

Python CLI

# Generate audio from text
python3 voice_generator.py tts --text "Hello, this is [PRINCIPAL_NAME]." --output hello.mp3

# Generate audio from script file
python3 voice_generator.py tts --script /workspace/voice/scripts/vsl.md

# Check configuration status
python3 voice_generator.py status

# List pending calls
python3 voice_generator.py calls --action list

# Weekly call summary
python3 voice_generator.py calls --action summary

Works Best Combined With

virtual-desktop → optional, for browser-based setup (see references/setup_guide.md)
acquisition-master → triggers calls when leads don't respond to email
funnel-builder → triggers calls when leads visit pricing page
content-creator-pro → provides scripts for audio content generation

Author

[PRINCIPAL_NAME]