Reader Podcast

A self-hosted web app that converts web articles into podcast episodes using Microsoft Edge TTS. Submit article URLs or subscribe to RSS feeds - the system scrapes the text, converts it to speech, and serves a per-user RSS podcast feed.

Features

• Submit article URLs and get audio versions in your podcast app
• Subscribe to RSS feeds for automatic article-to-audio conversion
• Per-user podcast feeds with API key authentication
• 47 English voices via Microsoft Edge TTS (no GPU required)
• Background processing with status tracking
• Admin endpoints for user management

Quick Start

Requirements

• Python 3.11+
• uv (Python package manager)
• ffmpeg (brew install ffmpeg or apt install ffmpeg)

Setup

git clone https://github.com/aedocw/reader-podcast.git
cd reader-podcast
uv sync

Configure

Create a .env file:

SITE_URL=https://pod.example.com   # Base URL for feed/mp3 links
ADMIN_KEY=your-secret-admin-key    # Required for /admin endpoints
PORT=8025                          # Optional (default: 8025)
DEFAULT_VOICE=en-US-AndrewNeural   # Optional (default: en-US-AndrewNeural)

Create a User

# Start the server
uv run python -m app.serve

# In another terminal, create a user
curl -X POST "http://localhost:8025/admin/users?key=your-secret-admin-key" \
  -d "username=myname"

This returns the user's api_key and feed_token.

Usage

• Add articles: Visit http://localhost:8025/add?key=YOUR_API_KEY
• RSS feed: Add http://localhost:8025/feed/YOUR_FEED_TOKEN to your podcast app
• Manage subscriptions: Visit http://localhost:8025/subscriptions?key=YOUR_API_KEY
• List voices: GET http://localhost:8025/voices

Docker

docker build -t reader-podcast .
docker run -p 8025:8025 --env-file .env -v reader-podcast-data:/app/data reader-podcast

Example docker-compose.yml:

services:
  reader-podcast:
    build: .
    container_name: reader-podcast
    env_file:
      - .env
    ports:
      - "8025:8025"
    volumes:
      - reader-podcast-data:/app/data
    restart: unless-stopped

volumes:
  reader-podcast-data:

Migration from Old Version

If you have an existing mp3/feed.xml from the previous Kokoro TTS version:

uv run python migrate.py --username admin

This creates a user and imports all existing episodes into the new SQLite database.

Environment Variables

Variable	Default	Description
SITE_URL	http://localhost:8025	Base URL for feed/mp3 links
ADMIN_KEY	(required)	Secret key for admin endpoints
PORT	8025	Server port
DATABASE_PATH	data/reader.db	SQLite database path
MP3_DIR	data/mp3	MP3 file storage directory
DEFAULT_VOICE	en-US-AndrewNeural	Default Edge TTS voice
RSS_POLL_INTERVAL_SECONDS	3600	How often to check RSS subscriptions

Project Structure

app/
  serve.py        - Bottle web routes
  db.py           - SQLite database layer
  tts.py          - Edge TTS synthesis
  scraper.py      - Article extraction
  feed_gen.py     - RSS feed generation
  rss_monitor.py  - RSS feed polling
  worker.py       - Background processing
  config.py       - Configuration
  auth.py         - Authentication
  text_clean.py   - Text normalization
templates/        - HTML templates