Voice & music
kesarAI can join a voice channel for two things: reading text aloud with text-to-speech, and playing music. Both run through the same voice connection — the bot holds one slot per server — and both are tuned per user on kesarai.ch while a server admin keeps final say over what’s allowed.
Overview
Text-to-speech and music are both on by default. Any member who is sitting in a voice channel can trigger TTS or queue music; the bot joins the channel of the member who asks. A server admin can switch text-to-speech off, or limit it to one role, on the voice dashboard. When nobody’s using it, the bot disconnects on its own — TTS leaves after about 60 seconds of silence, and the music player leaves after five minutes with nothing playing or with nobody listening in the channel.
Personal preferences live on your account pages and the server’s defaults live on the dashboard. Where the two disagree, the server can choose to force its settings over yours (see per-user vs server settings).
Text-to-speech
/ai tts makes the bot join your current voice channel and speak the text you give it aloud, using Google Cloud TTS. You have to be in a voice channel when you run it — that’s how the bot knows where to speak. After it finishes, the bot stays for about 60 seconds of silence and then leaves on its own. To send it out sooner, a member with Manage Channels can run /leavevc.
- Each call uses a little of your AI battery. Studio voices use about 10× as much, and if the bot owner has turned on low power, the standard voice speaks instead while your battery is low.
- You pick a preferred voice on your account page; if you don’t pass a voice to the command, that saved default is used.
- A server admin can lock a single default voice for everyone, which overrides individual preferences.
- Text is run through the same content filter as
/say, so banned words are censored before the bot speaks them.
On the voice dashboard, a server admin can set a required role: only members with that role can use TTS. Leave it empty and everyone is allowed. Create TTS role makes a role named “TTS” with no permissions, sets it as the required role and gives it to nobody — you hand it out to the members who should be allowed. It’s safe to click more than once: an existing role with that name is reused, not duplicated.
Which voice plays, in order: a voice named in the command, then the server voice if the server forces it (Force server voice), then the member’s own pick, then the server’s default voice, then kesarAI’s default. The cost follows whichever voice actually plays.
Toggle your personal TTS access and choose your voice at /account/voice.
Music playback
There are two ways to play music. Ask kesarAI in chat — ksr play <song>, ksr skip, ksr turn it down, ksr what’s this song — or use the /music commands. Asking works on every chat model and uses your AI battery like any other chat reply; /music uses none. Music works in servers only, and you join a voice channel first.
When a song starts, kesarAI posts a now-playing panel in the channel where the music was started. It shows the song, who asked for it and what’s next, with buttons for previous, pause, skip, stop, loop, shuffle, autoplay and the queue.
/music play <query>— play a song or playlist: a name or a link. kesarAI joins your voice channel; if a song is already playing, yours is added to the queue./music pause— pause or resume./music skip— skip to the next song./music previous— go back to the previous song./music stop— stop the music, clear the queue and leave the voice channel./music queue [page]— show what’s coming up, 10 songs per page./music nowplaying— post the now-playing panel in the channel you’re in./music autoplay [on|off]— keep playing similar songs when the queue runs out, or stop doing so. Leave the option empty to switch./music loop [off|this song|the queue]— loop the current song or the whole queue. Leave the option empty to step through the three./music shuffle— shuffle the queue once./music volume <0-150>— set the volume in percent; 100 is normal./music remove <position>— remove a song from the queue by its number in/music queue./music move <from> <to>— move a song to another place in the queue./music clear— clear the queue. The current song keeps playing./music seek <time>— jump to a time in the current song, like1:23or a number of seconds./music join— bring the bot into your current voice channel without playing anything./music status— check whether music is working: it shows the health of the music sources.
Autoplay. When the queue runs out, kesarAI keeps playing similar songs for as long as someone is listening. Autoplay is on by default; /music autoplay or the panel’s autoplay button switches it for the current session. A song you ask for replaces an autoplay song straight away.
Who can use the controls. To play a song or bring the bot in, you need to be in a voice channel. If kesarAI is already playing for people in another channel, your song is queued there and the reply says so. The other controls work for members in kesarAI’s voice channel; members with Manage Channels can use them from anywhere. Anyone can look at the queue and what’s playing.
Limits. A song can be up to 30 minutes long (radio and live streams aren’t limited) and each member can have 25 songs waiting in the queue.
Sources
The player resolves queries from four kinds of source. A bare search term is treated as a YouTube search; a pasted URL is routed by its domain.
- YouTube — videos, and playlists that expand into multiple queued tracks.
- SoundCloud — track and playlist links.
- Spotify — track, album, and playlist links, resolved to a matching audio stream under the hood.
- Internet radio — direct HTTP stream URLs, the most reliable source.
An admin chooses which of these sources are allowed on the server. A query that points at a disabled source is refused with a note to ask an admin.
Per-user vs server settings
Audio is configured at two levels. Each member can save their own volume, equaliser, and filter presets; these apply to the tracks they queue. The server holds its own defaults — a DJ role, a default volume, the allowed sources list, and a 24/7 stay-in-channel option that keeps the bot connected.
By default a member’s preferences win for their own tracks. When the server turns on force server audio, the server’s settings override every member’s personal audio instead.
Set your personal volume and EQ at /account/music.
Voice rooms
A hub is a voice channel. When a member joins it, kesarAI makes them their own voice room, moves them in, and posts a control panel in the room’s chat. The room is deleted once the last person leaves (after a short wait, so a dropped connection doesn’t end it). A server can have up to three hubs, each with its own room name, people limit and audio quality.
- The owner runs the room with the panel’s buttons: Rename, Limit, Lock, Hide, Permit, Remove and Transfer. There are no slash commands. Admins choose which buttons owners may use, per hub.
- If the owner leaves while others stay, a Claim button opens after a minute so someone inside can take over.
- Rooms copy their category’s permissions, so jail, quarantine, mutes and verification still apply. Owners never get Discord permissions on the channel; everything goes through the buttons.
- Staff aren’t let into locked or hidden rooms automatically: access follows the channel and category permissions. Each hub has an optional Always let in list of roles that can always see and join its rooms.
kesarAI needs Manage Channels, Move Members and Manage Roles in the rooms’ category; the dashboard warns when one is missing. If the server uses another security bot, whitelist kesarAI there so it isn’t flagged for making channels. Set it up at Voice rooms on the dashboard.
Command reference
Text-to-speech
/ai tts— speak text in your voice channel./leavevc— make the bot leave the voice channel it joined for text-to-speech. Needs Manage Channels.
Music
/music play,/music pause,/music skip,/music previous,/music stop,/music queue,/music nowplaying,/music autoplay,/music loop,/music shuffle,/music volume,/music remove,/music move,/music clear,/music seek,/music join,/music status.- Or ask in chat:
ksr play <song>,ksr skip,ksr pause.
/music stop ends music; /leavevc only ends text-to-speech. Each command is also listed in the slash commands reference.
