🔊 Guide

Get Bozos TTS speaking naturally

Start with a clean voice connection, then configure the current voice channel, personal overrides, voice previews, optional Spatial Speakers and Auto-Join.

1. Join the voice channel first

Enter the Discord voice channel you want to use and run /join. Bozos connects to your current VC and begins handling eligible text-channel messages for that voice session.

If the bot does not join, confirm that you are in a VC and that Bozos TTS can View Channel, Connect and Speak there. Channel/category overrides can block access even when the bot role looks correct.

2. Choose the correct language scope

Run /language and choose either Voice channel or Only me. Voice channel changes the default for that specific VC. Only me stores a personal preference that takes priority for your own messages.

3. Pick a regional locale and voice

The production catalog contains 75 language families, 142 regional locales and 323 verified neural voices. After choosing the language/accent, use /voice with the same Voice channel or Only me scope.

4. Preview voices privately

Run /voicepreview while you are in the same active VC as Bozos. The preview panel is private and sends preview audio as Discord voice messages, so you can compare voices without interrupting the live TTS queue.

If previews do not send, check Attach Files and Send Voice Messages in the text channel where you ran the command.

5. Let pronunciation and chat cleanup work automatically

The Grand Pronunciation Engine improves common slang, acronyms, gaming/tech terms, file formats, units and Discord-style usernames. The speech pipeline also normalizes markdown, mentions, URLs, emoji, numbers, dates, times, punctuation and expressive chat spelling.

6. Choose a speaker identity system

The default Username System intelligently announces names only when useful. To avoid spoken usernames, use /speaker-system set and choose Spatial Speakers. Each member then receives a stable spatial position plus a compatible voice; a personal /voice selection still wins.

Use /speaker-system map to view current assignments and /speaker-system test for a Left → Center → Right headphone test.

7. Adjust live playback

/volume sets playback from 0% to 200% for the current VC and requires Manage Server. The change applies immediately even if speech is already playing. /skip skips the current message and continues the queue.

8. Optional join/leave announcements

A member with Manage Server can use /announce in the same active VC to enable or disable localized join/leave announcements for that voice channel. They are off by default.

9. Optional persistent Auto-Join

To configure Auto-Join, first connect Bozos normally with /join, then run /auto-join enabled:true from that same VC. You need Manage Server.

Auto-Join requires View Channel, Connect, Speak and Send Messages in the configured VC. At least one person in that VC must be a member of Bozos | Official Community for Auto-Join to operate. Use /auto-join enabled:false to disable the persistent configuration; an already-running session is not forcibly ended.

10. Multiple voice channels in one server

Bozos can coordinate the main TTS bot with two voice-only worker identities. When those workers are installed and available, up to three independent TTS sessions can run in different VCs of the same Discord server.

11. When speech does not play

  1. Confirm you and Bozos are in the expected voice channel.
  2. Check View Channel, Connect and Speak permissions.
  3. Confirm the message is being sent in the text channel associated with the active voice session.
  4. Try a short plain-text message.
  5. If the session disconnected, reconnect with /join.
  6. If only one voice fails, try another voice in the same locale; Bozos can also fall back to another compatible voice at runtime.
Full TTS documentation →TTS commands →