The latest updates on your projects. Learn more about Vercel for GitHub.

2 Skipped Deployments

Project	Deployment	Actions	Updated (UTC)
discord-js	Skipped		Mar 4, 2026 2:01pm
discord-js-guide	Skipped		Mar 4, 2026 2:01pm

sequenceDiagram
    actor UDP as UDP Source
    participant VR as VoiceReceiver
    participant CAP as createAudioPacket
    participant ARS as AudioReceiveStream
    
    UDP->>VR: onUdpMessage (encrypted packet)
    Note over VR: Extract RTP sequence,<br/>timestamp from bytes
    VR->>VR: Decrypt packet payload
    VR->>CAP: createAudioPacket(buffer,<br/>sequence, timestamp, ssrc)
    CAP->>CAP: Attach metadata as<br/>non-enumerable properties
    CAP-->>VR: AudioPacket (Buffer + metadata)
    VR->>ARS: stream.push(AudioPacket)
    Note over ARS: Consumer receives<br/>Buffer with metadata

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

Codecov Report

❌ Patch coverage is 89.28571% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 31.70%. Comparing base (0cb8be4) to head (fef397d).
⚠️ Report is 10 commits behind head on main.

@@            Coverage Diff             @@
##             main   #11432      +/-   ##
==========================================
+ Coverage   31.66%   31.70%   +0.03%     
==========================================
  Files         386      386              
  Lines       13966    13977      +11     
  Branches     1098     1099       +1     
==========================================
+ Hits         4422     4431       +9     
- Misses       9410     9412       +2     
  Partials      134      134              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

FYI: code coverage fail is misleading, this is existing lack of coverage (these lines are just reorganisations of the code that was already there to make it easier to understand). It looks like there's no test data for DAVE

For testing sake, can you draft up a quick code sample that can be used to see how it works now? (bonus point, we can document the breaking change)

@vladfrangu Sure, here's a sample snippet that'll join a voice channel and subscribe to all speakers, is this what you were thinking of?

import type { VoiceBasedChannel } from 'discord.js';
import {
	joinVoiceChannel,
	entersState,
	VoiceConnectionStatus,
	EndBehaviorType,
	type AudioPacket,
} from '@discordjs/voice';

async function joinAndRecordChannel(channel: VoiceBasedChannel) {
	// Join the voice channel
	const connection = joinVoiceChannel({
		channelId: channel.id,
		guildId: channel.guild.id,
		adapterCreator: channel.guild.voiceAdapterCreator,
		selfDeaf: false, // We want to receive audio
		selfMute: true, // We don't want to send audio
	});

	// Wait for the connection to be ready
	await entersState(connection, VoiceConnectionStatus.Ready, 30_000);

	const receiver = connection.receiver;

	// Subscribe to all users when they speak
	const subscribedUsers = new Set<string>();
	receiver.speaking.on('start', (userId) => {
		if (subscribedUsers.has(userId)) return;
		subscribedUsers.add(userId);

		// Subscribe to the user's audio stream
		const audioStream = receiver.subscribe(userId, {
			end: { behavior: EndBehaviorType.Manual },
		});

		// New API: 'data' event now emits an AudioPacket rather than a Buffer
		audioStream.on('data', (packet: AudioPacket) => handleOpusPacket(userId, packet));

		audioStream.on('error', (err) => {
			console.error(`Audio stream error for ${userId}:`, err);
		});
	});
}

function handleOpusPacket(userId: string, packet: AudioPacket) {
	console.log(
		`User ${userId}: rtp.ssrc=${packet.ssrc} rtp.seq=${packet.sequence} rtp.ts=${packet.timestamp}@48kHz opus=${packet.payload.length}B`,
	);
}

(a better example might show how to use the new fields in opus decoding to produce a fully synchronised PCM from all speakers, but I've extracted this from my transcription bot which defers opus decode to an offline process to minimise load while recording)

Here's an initial draft of the breaking change doc entry (have tried to match general style from the Changes in v14 .mdx):

AudioReceiveStream

AudioReceiveStream now emits AudioPacket instances instead of raw Buffers. Each AudioPacket contains the Opus payload along with relevant RTP header metadata (sequence, timestamp, and ssrc).

The Opus audio data can be found in the payload property:

- // Previously, each chunk was a raw Buffer of Opus data:
- receiver.subscribe(userId).on('data', (chunk) => {
- 	processOpusData(userId, chunk); // chunk was a Buffer
- });

+ // Now, each chunk is an AudioPacket with payload and RTP metadata:
+ receiver.subscribe(userId).on('data', (packet: AudioPacket) => {
+ 	processOpusData(userId, packet.payload); // Opus data is on .payload
+ 	console.log(packet.sequence);    // RTP sequence number
+ 	console.log(packet.timestamp);   // RTP timestamp (48kHz clock)
+ 	console.log(packet.ssrc);        // RTP synchronization source
+ });

That is perfect, thanks!