yt-pod-rss

Turn a YouTube channel into a podcast RSS feed with downloadable audio files.

How it works

1. Fetches the latest videos from a YouTube channel (skips Shorts)
2. Downloads each video's audio as an .m4a file into media/
3. Generates an RSS 2.0 feed (feed.xml) with <enclosure> elements pointing to the local audio files
4. Can run locally or via GitHub Actions (scheduled daily) and deploy to GitHub Pages

Quick start — GitHub Actions (hosted)

This is the easiest way: fork the repo, configure it, and GitHub will regenerate and host your feed daily.

1. Fork

Fork this repo on GitHub.

2. Set your channel URL

Edit config.yml in your fork and add your YouTube channels:

feeds:
  - name: my-podcast
    channel_url: https://www.youtube.com/@YourChannel

3. Export YouTube cookies

Follow the yt-dlp guide on exporting YouTube cookies to get a cookies.txt file.

4. Add cookies to your fork

You have two options:

Option A — GitHub secret (safer, recommended):

1. Open your fork's Settings → Secrets and variables → Actions
2. Create a new repository secret named YT_COOKIES
3. Paste the entire contents of your cookies.txt file as the value

Option B — Commit cookies.txt (simpler, private forks only):

1. Copy the exported cookies.txt into the repo root
2. It's already in .gitignore — uncomment or remove it from .gitignore if you want it committed

5. Enable GitHub Pages

1. Go to your fork's Settings → Pages
2. Under Source, select GitHub Actions
3. The feed will be available at https://<your-username>.github.io/yt-pod-rss/feed.xml

6. Run it

• The workflow runs automatically every day at 06:00 UTC
• You can also trigger it manually: Actions → Generate RSS Feed → Run workflow
• After the first run, add the RSS feed URL to your podcast app

Running locally

# Install dependencies
pip install -r requirements.txt

# Install Deno (required by yt-dlp for JS challenge solving)
curl -fsSL https://deno.land/install.sh | sh
export PATH="$HOME/.deno/bin:$PATH"

# Generate the feed (uses channel_url from config.yml)
python generate_feed.py

# Or specify a URL directly
python generate_feed.py "https://www.youtube.com/@channel/videos"

# Test mode (uses sample data, no network calls)
python generate_feed.py --test

# Custom number of videos
python generate_feed.py -n 10

If you get bot-blocked, place a cookies.txt file in the repo root or set the YT_COOKIES environment variable.

Running with Docker

# Build the image
docker build -t yt-pod-rss .

# Run with config.yml defaults (mount media/ and cookies.txt for persistence)
docker run -v "$PWD/media:/app/media" -v "$PWD/cookies.txt:/app/cookies.txt" yt-pod-rss

# Pass arguments directly to generate_feed.py
docker run -v "$PWD/media:/app/media" yt-pod-rss --test

# Single URL mode (override config)
docker run -v "$PWD/media:/app/media" yt-pod-rss "https://www.youtube.com/@channel" -n 5

# Use YT_COOKIES env var instead of a file
docker run -e YT_COOKIES="$(cat cookies.txt)" -v "$PWD/media:/app/media" yt-pod-rss

Project structure

yt-pod-rss/
├── config.yml                 # Feed list (name + channel_url per feed)
├── generate_feed.py           # Main script
├── requirements.txt           # Python dependencies
├── test_data.json             # Sample data for --test mode
├── cookies.txt                # YouTube cookies (optional, not committed)
├── Dockerfile                 # Docker image definition
├── media/                     # Downloaded audio files (not committed)
├── *.xml                      # Generated RSS feeds (not committed)
└── .github/workflows/
    └── generate.yml           # GitHub Actions workflow (daily schedule)

---

Built by @andreparames with OpenCode