Skip to content

Quick Start

Get your first AI-generated podcast episode in about 5 minutes.

Prerequisites

What you need

Required

  • Anthropic API key — Claude script generation. Get one at console.anthropic.com.
  • Google Cloud project with TTS + GCS enabled. Use the Terraform setup to provision everything automatically.

Optional

AWS S3 (hosting), MongoDB (deduplication), and Grafana Loki (logging) are all optional. The core pipeline works without them — output is saved locally.

5-minute quickstart

Step 1 — Pull the image

docker pull ghcr.io/thewintershadow/the-data-packet:latest

Step 2 — Create an environment file

cat > .env << 'EOF'
ANTHROPIC_API_KEY=sk-ant-api03-your-key-here
GCS_BUCKET_NAME=your-gcs-bucket-name
GOOGLE_APPLICATION_CREDENTIALS=/credentials.json
EOF

Step 3 — Fix output directory permissions

mkdir -p output
sudo chown -R 1000:1000 output  # (1)!
  1. The container runs as UID 1000. This ensures it can write to the mounted volume.

Step 4 — Run

docker run --rm \
  --env-file .env \
  -v "$(pwd)/output:/app/output" \
  -v "$(pwd)/service-account-key.json:/credentials.json:ro" \
  ghcr.io/thewintershadow/the-data-packet:latest

pip install the-data-packet

# Set required env vars
export ANTHROPIC_API_KEY="sk-ant-api03-your-key"
export GCS_BUCKET_NAME="your-gcs-bucket"
export GOOGLE_APPLICATION_CREDENTIALS="./service-account-key.json"

# Run
the-data-packet --output ./episode

What you'll get

After a few minutes the pipeline completes and you'll have:

output/
├── episode_script.txt    # Two-host dialogue script
└── episode.wav           # Synthesized audio file

That's it!

You now have a complete podcast episode. The script is the raw dialogue, and the .wav file is ready to upload to any podcast host.

What just happened?

sequenceDiagram
    participant You
    participant Pipeline
    participant Wired
    participant Claude
    participant GoogleTTS

    You->>Pipeline: docker run ...
    Pipeline->>Wired: fetch RSS feed
    Wired-->>Pipeline: latest articles
    Pipeline->>Claude: write podcast script
    Claude-->>Pipeline: two-host dialogue
    Pipeline->>GoogleTTS: synthesize audio
    GoogleTTS-->>Pipeline: episode.wav
    Pipeline-->>You: output/episode_script.txt + episode.wav

Next steps

  • Installation options

    All ways to install: Docker, pip, or from source, with platform notes.

    Installation →

  • CLI reference

    Every flag, voice option, source selector, and output setting documented.

    Usage →

  • Configuration

    All environment variables with defaults and examples.

    Configuration →