# Jarri Spooty

Self-hosted Spotify playlist and track downloader built with NestJS + Angular.

Jarri Spooty does not download audio from Spotify itself.  
It retrieves metadata from Spotify and locates matching audio on YouTube.

![Jarri Spooty Demo](docs/images/jarri-spooty-demo.png)

Deterministic Spotify metadata ingestion with duration-aware YouTube candidate scoring, paced queue execution, hardened Docker deployment, and local archival workflows.

This hardened branch focuses on:

- Large playlist support (>100 tracks)
- Spotify OAuth login flow
- Better YouTube pacing and throttling resistance
- Improved Docker deployment
- Safer credential handling
- More resilient cover-art embedding
- Better queue stability
- Deterministic YouTube fallback handling
- Explicit yt-dlp CLI execution
- Automatic failed-candidate rejection
- Improved age-gated video handling
- Persistent SQLite state across container restarts
- Deterministic Docker config persistence
- Improved operational observability
- Reduced unused dependency surface
- Hardened filename, cover-art, subprocess, and websocket boundaries

---

# Features

- Download Spotify playlists
- Download individual Spotify tracks
- Playlist auto-subscription support
- Automatic YouTube matching
- MP3 tagging and embedded cover art
- Docker deployment
- Queue-based download system
- Spotify OAuth integration
- Large playlist pagination support
- Download pacing controls
- YouTube cookie support
- Automatic YouTube retry/fallback handling
- Failed YouTube candidate rejection memory
- Deterministic yt-dlp error classification
- Hardened cover-art validation and embedding
- Explicit client-facing websocket payload shaping

---

# Important Notice

Use this software responsibly.

Only download music you legally own or are permitted to access.

The maintainers are not responsible for misuse.

---

# Supported URLs

- Spotify playlists
- Spotify tracks

Example:

    https://open.spotify.com/playlist/...
    https://open.spotify.com/track/...

---

# Quick Start (Recommended)

## 1. Create Spotify Developer App

Go to:

    https://developer.spotify.com/dashboard

Create an application.

Add this Redirect URI:

    http://127.0.0.1:3000/api/spotify/callback

Copy:

- Client ID
- Client Secret

---

## 2. Store Credentials Outside Repository

Create a secure env file:

~~~bash
sudo mkdir -p /etc/tokens

sudo tee /etc/tokens/spotify.env > /dev/null <<'EOT'
SPOTIFY_CLIENT_ID=your_client_id
SPOTIFY_CLIENT_SECRET=your_client_secret
EOT

sudo chown root:$USER /etc/tokens/spotify.env
sudo chmod 640 /etc/tokens/spotify.env
~~~

Never commit this file.

---

## 3. Export YouTube Cookies (Recommended)

YouTube increasingly rate-limits or age-gates anonymous downloads.

Export a Netscape-format `cookies.txt` from a logged-in browser session.

Recommended storage:

~~~bash
sudo cp cookies.txt /etc/tokens/youtube.cookies.txt
sudo chown root:$USER /etc/tokens/youtube.cookies.txt
sudo chmod 640 /etc/tokens/youtube.cookies.txt
~~~

---

# Docker Run

~~~bash
docker run --rm -p 3000:3000 \
  --env-file /etc/tokens/spotify.env \
  -e SPOTIFY_REDIRECT_URI='http://127.0.0.1:3000/api/spotify/callback' \
  -e AUTH_ENABLED=true \
  -e SPOOTY_AUTH_TOKEN=change_this_token \
  -e YT_SEARCH_DELAY_MS=7000 \
  -e YT_DOWNLOADS_PER_MINUTE=6 \
  -e YT_COOKIES_FILE=/spooty/config/youtube.cookies.txt \
  -v "$PWD/downloads:/spooty/backend/downloads" \
  -v "$PWD/spooty-config:/spooty/backend/config" \
  -v "/etc/tokens/youtube.cookies.txt:/spooty/config/youtube.cookies.txt:ro" \
  jarri-spooty:local
~~~

Open:

    http://127.0.0.1:3000/?token=change_this_token

Then:

1. Click "Connect Spotify"
2. Login to Spotify
3. Approve access
4. Paste playlist URL
5. Download

---

# Deterministic YouTube Fallback Handling

The hardened branch now uses direct `yt-dlp` CLI execution rather than relying entirely on wrapper abstractions.

This provides:

- Explicit stderr visibility
- Better Docker compatibility
- Deterministic retry handling
- Automatic failed-candidate rejection
- Better age-gated video handling
- Improved operational observability

If a YouTube candidate fails:

1. The failed URL is recorded
2. The candidate is rejected
3. A new YouTube search is performed
4. The next-best valid candidate is attempted automatically

This prevents infinite retry loops against dead or restricted videos.

---

# Queue Pacing

Aggressive YouTube access can trigger:

- HTTP 302 loops
- CAPTCHA
- temporary throttling
- incomplete downloads

Recommended safe pacing:

~~~bash
-e YT_SEARCH_DELAY_MS=7000
-e YT_DOWNLOADS_PER_MINUTE=6
~~~

The hardened branch also includes additional internal pacing and retry coordination to reduce:

- repeated failed candidate loops
- aggressive retry bursts
- queue collisions
- YouTube anti-bot triggers

---

# Security Notes

Never commit:

- Spotify secrets
- OAuth tokens
- cookies.txt
- downloaded music
- local databases
- spooty-config/

Recommended `.gitignore` additions:

~~~gitignore
downloads/
config/
spooty-config/
*.sqlite
cookies.txt
.env
.env.local
~~~

---

# License

MIT