DodoBot

A modern Discord music bot. YouTube + Spotify, slash commands, rich embeds with interactive controls, and auto-recommend. Designed to run 24/7 on an AWS t3.micro EC2 free-tier instance.

Rewritten in 2026: migrated off the unmaintained pytube to yt-dlp 2026.3+, discord.py 2.7+, spotipy. Speech-recognition features removed.

Features

• /play with a multi-result picker - if your query matches several tracks, a dropdown appears so you can pick the right one before anything is enqueued.
• Rich embeds for every response: now-playing shows title, thumbnail, duration, uploader, queue length, and auto-recommend state.
• Interactive control panel attached to the now-playing message: Pause/Resume, Skip, Shuffle, Auto-recommend toggle, Refresh, Disconnect (discord buttons).
• Auto-recommend mode: when the queue empties and auto-recommend is ON, the bot pulls the YouTube Mix of the last track and keeps playing related songs.
• Direct URL support: paste any YouTube video / playlist / shorts link, or any Spotify track / album / playlist / artist link, and it just works. No picker shown for direct URLs (it goes straight into the queue).
• Streaming, not downloading: yt-dlp extracts the stream URL and FFmpeg pipes it directly to Discord - no temp files written to disk.

Slash commands

Command	Description
/play query:<text or URL>	Enqueue a song. Text queries show a 5-result picker; URLs enqueue directly.
/pause	Pause or resume playback.
/skip	Skip the current track.
/queue	Show the upcoming queue.
/shuffle	Shuffle the queue.
/clear	Clear the queue and stop playback.
/autorecommend	Toggle auto-recommend mode.
/disconnect	Disconnect the bot from voice.

Environment variables

Copy .env.example to .env and fill in:

DISCORD_TOKEN=<your bot token>
SPOTIFY_CLIENT_ID=<optional>
SPOTIFY_CLIENT_SECRET=<optional>

• DISCORD_TOKEN (required) - the bot token from the Discord developer portal.
• SPOTIFY_CLIENT_ID / SPOTIFY_CLIENT_SECRET (optional) - Spotify Web API app credentials. Without them the bot still plays YouTube fine; Spotify URLs will return a friendly error.
  • Heads up: as of 2026, Spotify restricts the Client Credentials flow to apps whose owner has an active Spotify Premium subscription. If you hit 403 premium required, upgrade the app-owner account.

Legacy names TOKEN, SPOTIFY_ID, SPOTIFY_SECRET are still honoured so existing .env files keep working.

Local run

Requires Python 3.11+ and ffmpeg on the PATH.

python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt
Copy-Item .env.example .env   # then edit .env
python -m bot.main

On Linux/macOS:

python3.12 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env          # then edit .env
python -m bot.main

Tests

pip install -r requirements-dev.txt
python -m pytest                    # unit tests only
python -m pytest -m integration     # live yt-dlp + spotify checks
python -m ruff check .

AWS deployment (free tier)

Why EC2 and not Lambda? A Discord music bot needs a persistent WebSocket gateway and a UDP voice connection. AWS Lambda has a 15-minute execution cap and charges per millisecond - a 24/7 WebSocket cannot run there, and Lambda does not support UDP at all. Lambda is only useful for stateless HTTP-Interactions bots (no voice).

So the deployment target is a single t3.micro EC2 instance (AWS 12-month free tier: 750 hours/month - exactly enough for one instance running 24/7). After the 12-month window it costs ~$7.50/month, or drop to t4g.nano (ARM Graviton) for ~$3/month.

Provision with one command

From a machine with aws-cli authenticated:

.\deploy\launch.ps1 -Region us-east-1 -RepoUrl https://github.com/<you>/<repo>.git

This:

1. Creates a key pair dodobot-key (PEM saved to deploy\dodobot-key.pem, locked down via icacls).
2. Creates a dodobot-sg security group that only allows SSH from your current public IP.
3. Finds the latest Amazon Linux 2023 AMI.
4. Launches a t3.micro with 30 GB gp3, passing deploy/bootstrap.sh as user-data.

bootstrap.sh runs on first boot and:

• installs Python 3.12, git, and a static ffmpeg build,
• creates the dodobot system user,
• clones this repo to /opt/dodobot,
• creates a venv and installs requirements.txt,
• installs deploy/dodobot.service and enables it.

Upload your .env

The service won't start until DISCORD_TOKEN is populated.

scp -i .\deploy\dodobot-key.pem .\.env ec2-user@<public-ip>:/tmp/.env
ssh -i .\deploy\dodobot-key.pem ec2-user@<public-ip> `
  'sudo mv /tmp/.env /opt/dodobot/.env && sudo chown dodobot:dodobot /opt/dodobot/.env && sudo chmod 600 /opt/dodobot/.env && sudo systemctl restart dodobot'

Logs

ssh -i .\deploy\dodobot-key.pem ec2-user@<public-ip> 'sudo journalctl -u dodobot -f'

Updating

SSH in and:

cd /opt/dodobot
sudo -u dodobot git pull
sudo -u dodobot .venv/bin/pip install -r requirements.txt
sudo systemctl restart dodobot

Project layout

bot/
  main.py           # entry; builds bot, syncs slash commands
  config.py         # env loading
  audio.py          # yt-dlp streaming + FFmpeg source + YouTube Mix fetch
  search.py         # URL routing (YouTube / Spotify / text query)
  queue.py          # per-guild GuildQueue
  player.py         # playback loop, voice client, auto-recommend
  embeds.py         # rich-embed builders
  spotify_client.py # spotipy wrapper: URLs + search
  cogs/
    music.py        # Music cog (slash commands)
    views.py        # SongPickerView + ControlPanelView
deploy/
  launch.ps1        # AWS CLI provisioning
  bootstrap.sh      # EC2 user-data
  dodobot.service   # systemd unit
tests/              # pytest suite

What changed from v1

• Removed: speech_recognition, pyttsx3, pyaudio, gtts, and all voice-command code.
• Removed: 82 MB ffmpeg.exe Windows binary from the repo (installed on-host instead).
• Removed: [prefix] message commands ([radio], [sradio], [pause], etc.) in favour of slash commands.
• Removed: pytube (unmaintained; YouTube detection loops broke).
• Removed: download-to-disk playback (<title>.mp3 files). Everything streams now.
• Added: multi-result picker, interactive button panel, rich embeds, auto-recommend.
• Added: systemd service, EC2 bootstrap scripts, AWS CLI one-shot launcher.

License

See LICENSE in the repo.