Skip to content
An open-source Java Teamspeak 3 client library using the TS3 full client protocol
Branch: master
Clone or download
Pull request Compare This branch is 11 commits behind Manevolent:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.



TS3J is an open-source implementation of the reverse-engineered Teamspeak3 full server/client protocol, as an adaptation of Splamy's C# TS3Client source code. You can find that here:

This project is the network-level API library for the reference implementation of Teamspeak3 for Manebot, my multiplatform chatbot that you can extend with your own plugins. GitHub:

A standalone proof-of-concept was created to wrap ts3j:

(Shameless plug) If you want to deliver a great music bot, check out ffmpeg4j, a wrapper around the native C library FFmpeg, to get insanely low CPU usage on your YouTube video downloads (or streams if you don't want to hit disk), and audio playback:


If you want the latest -SNAPSHOT:


Connection & Basic Setup

client = new LocalTeamspeakClientSocket();

// Set up client

   new InetSocketAddress(
        port // UDP client port, Teamspeak3 client uses 9987

// Subscribe to all channels

// Get list of clients
for (Client basicClient : client.listClients())

Note that while connect() is processing, you'll receive channels registered and clients currently connected to the server. It is important that you use the listener to collect these, and track their changes through the other listener event calls.

You can interact with the server using the commands on the client object similarly to TS3Query.

Handling chat

// TS3Listener interface
public void onTextMessage(TextMessageEvent textMessageEvent) {
	if (textMessageEvent.getInvokerId() == client.getClientId())

	// Global chat example
	// PM to the sender
	client.sendPrivateMessage(textMessageEvent.getInvokerId(), "Echo!");

Sending audio

// Microphone interface
public void write(float[] buffer, int len) {
	byte[] opusPacket = doOpusEncodingHere(buffer, len);


public boolean isReady() {
	return true;

public CodecType getCodec() {
	return CodecType.OPUS_MUSIC;

public byte[] provide() {
	try {
	    if (packetQueue.peek() == null)
		return new byte[0]; // underflow

	    OpusPacket packet = packetQueue.remove();

	    if (packet == null)
		return new byte[0]; // underflow

	    return packet.getBytes();
	} catch (NoSuchElementException ex) {
	    return new byte[0]; // signals the decoder on the clients to stop

Receiving audio

Refer to the setVoiceHandler and setWhisperHandler methods to supply a Consumer object to receive Voice and Whisper packets. You will need to decode the packets yourself, and insert packet-loss-correction as needed.

Note that the first 5 packets starting a voice session are marked with the COMPRESSED flag. The final voice packet, intended to singal to close your decoder and flush samples, is always empty (0-length byte array).

Manebot can do this in its TS3 plugin, which uses TS3j:

You can’t perform that action at this time.