🎙️ AI Podcast Pipeline

Two-person dialogue podcast generator with natural conversations and emotions.

Use AI responsibly.

---

⚡ Quick Start

# 1. Install
pip install -r requirements.txt

# 2. Configure
cp config/.env.template config/.env
# Add your API keys

# 3. Run
python podcast_pipeline.py

---

🎯 Features

Natural Dialogue

• ✅ Provider-specific emotion tags (auto-selected per provider)
• ✅ Reactions ([laughs], [sighs], [gasps])
• ✅ Cartesia 5 base emotions + intensity, full ElevenLabs dynamics

Two TTS Providers

• Cartesia - sonic-3 model, API emotion controls (positivity:high, curiosity:medium, etc.)
• ElevenLabs - Premium voices, interruptions & overlapping speech

Scalable Architecture

• Provider config files (providers/configs/*.yaml)
• Template placeholders for easy 3rd provider addition
• Per-voice speed settings

Post-Processing Tools

• tune_audio.py - Adjust speaker speeds post-generation
• translate_script.py - Translate to other languages

---

🎛️ Primary Workflow

python podcast_pipeline.py

Prompts: Project name → Topic → Duration → Style → Language → Provider → Mode

Output:

projects/myprojectname/
├── scripts/myprojectname_DE_2025-12-06_14-30_CRTS_draft1.txt
├── audio/myprojectname_de_2025-12-06_CRTS_PRODUCTION.mp3
└── sources/myprojectname_sources.txt

---

📁 Project Structure

myfirstpodcast/
├── config/
│   ├── .env                    # API keys (gitignored)
│   └── podcast_config.json     # Voice IDs, speeds, styles
├── templates/                   # Script templates (3 styles × 3 languages)
├── providers/
│   ├── configs/                # Provider-specific emotion configs
│   │   ├── cartesia.yaml
│   │   └── elevenlabs.yaml
│   ├── cartesia.py
│   └── elevenlabs.py
├── projects/                    # Your podcasts (gitignored)
├── podcast_pipeline.py         # Main generator
├── tune_audio.py              # Speed adjustment
└── translate_script.py        # Script translation

---

🔧 Configuration

API Keys (config/.env)

ANTHROPIC_API_KEY=sk-ant-api03-...
ELEVENLABS_API_KEY=sk_...
CARTESIA_API_KEY=...

Voice IDs (config/podcast_config.json)

• 2 providers × 3 languages × 2 speakers = 12 voice IDs
• Optional: Set per-voice default speeds

---

🛠️ Tools Reference

Tool	Purpose
podcast_pipeline.py	Generate podcasts
tune_audio.py	Adjust speaker speeds
translate_script.py	Translate scripts
smart_update.py	Safe updates with backups

---

📄 License

CC BY-NC-SA 4.0 - Attribution, NonCommercial, ShareAlike

---

Generate natural, engaging podcasts in minutes! 🎙️