GoogleMeetInput

Defined in: src/providers/io/meet/GoogleMeetInput.ts:419

Browser audio input provider that joins a Google Meet conference via the Meet Media API (Developer Preview) and emits the meeting’s mixed audio.

Remarks

GoogleMeetInput negotiates a WebRTC session directly with Google Meet:

1. Creates an RTCPeerConnection (Google STUN, max-bundle).
2. Adds exactly AUDIO_TRANSCEIVER_COUNT | 3 recvonly audio transceivers (Meet uses 3 virtual audio SSRCs; Opus is negotiated by default).
3. Opens the required session-control and media-stats data channels (both ordered: true) before creating the SDP offer.
4. POSTs the offer to {endpoint}/v2beta/{spaceName}:connectActiveConference with a Bearer OAuth token and applies the returned SDP answer.
5. Mixes the incoming audio tracks through a shared AudioContext processing node (AudioWorklet preferred, ScriptProcessorNode fallback), downsamples to 16 kHz mono linear16, and emits AudioChunk objects while active.

The media-stats channel is serviced per the Meet reference client: the provider sends nothing until the server’s configuration resource arrives, then uploads allowlisted RTCPeerConnection.getStats() sections at the server-specified interval.

The Meet Media API is receive-only — this provider covers only the 'input' role and cannot inject audio into the conference.

Data-flow diagram:

Meet SFU ──RTP (Opus ×3)──▶ RTCPeerConnection.ontrack
                                   │
                    MediaStream ──▶ AudioContext (auto-mixed)
                                   │
                     worklet / script-processor (Float32)
                                   │
                 downsample 16 kHz → floatTo16BitPCM
                                   │
             active && !paused ? callback(chunk) : drop

Example

import { GoogleMeetInput } from 'composite-voice';

const meet = new GoogleMeetInput({
  apiKey: 'ya29.a0...', // OAuth token with meetings.conference.media.audio.readonly
  spaceName: 'spaces/jQCFfuBOdN5z',
  debug: true,
});

meet.onSessionStatus((status) => console.log('Meet session:', status.connectionState));
meet.onAudio((chunk) => console.log(`audio: ${chunk.data.byteLength} bytes`));

await meet.initialize(); // joins the conference
meet.start();            // begin emitting audio chunks
// ...
await meet.dispose();    // sends a leave request and tears down WebRTC

See

• GoogleMeetInputConfig for configuration options
• AudioInputProvider for the interface contract
• GoogleMeetSessionStatus for session lifecycle notifications

Implements

• AudioInputProvider

Constructors

Constructor

new GoogleMeetInput(config): GoogleMeetInput;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:522

Creates a new GoogleMeetInput instance.

Parameters

Returns

GoogleMeetInput

Remarks

Construction is side-effect free; the WebRTC session is established in initialize().

Example

const meet = new GoogleMeetInput({
  apiKey: async () => getFreshAccessToken(),
  spaceName: 'spaces/jQCFfuBOdN5z',
});

Properties

Methods

dispose()

dispose(): Promise<void>;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:644

Leave the conference and release all resources.

Returns

Promise<void>

Remarks

Sends a leave request on the session-control channel (per the Meet Media API contract: {"request":{"requestId":n,"leave":{}}}), then closes the data channels, the peer connection, and the AudioContext. The provider may be re-initialized afterwards. Safe to call multiple times.

Implementation of

AudioInputProvider.dispose

getMetadata()

getMetadata(): AudioMetadata;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:750

Get the audio format metadata for the emitted audio.

Returns

AudioMetadata

AudioMetadata — linear16, 16000 Hz, 1 channel, 16-bit.

Remarks

The provider always emits 16 kHz mono 16-bit linear PCM regardless of the AudioContext hardware rate (audio is downsampled internally). Used by the pipeline to auto-configure the STT provider.

Implementation of

AudioInputProvider.getMetadata

getSessionStatus()

getSessionStatus(): 
  | GoogleMeetSessionStatus
  | null;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:792

Get the most recent session status received from the server.

Returns

| GoogleMeetSessionStatus | null

The latest GoogleMeetSessionStatus, or null if no status has been received yet.

initialize()

initialize(): Promise<void>;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:549

Join the active conference: negotiate the WebRTC session with the Meet Media API and set up the audio mixing pipeline.

Returns

Promise<void>

Remarks

Performs the full connect flow described in the class documentation. The conference identified by GoogleMeetInputConfig.spaceName must be active, and the OAuth token must carry a Meet Media API audio scope. If already initialized, this is a no-op.

Throws

ProviderInitializationError when configuration is invalid, RTCPeerConnection is unavailable (non-browser environment), or WebRTC setup fails.

Throws

ProviderConnectionError when the connectActiveConference request is rejected (non-2xx response or a response without an SDP answer).

Implementation of

AudioInputProvider.initialize

isActive()

isActive(): boolean;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:722

Check whether the provider is actively emitting audio.

Returns

boolean

true when started and not paused.

Implementation of

AudioInputProvider.isActive

isReady()

isReady(): boolean;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:662

Check whether the provider is connected to the conference.

Returns

boolean

true after initialize completes and until dispose is called or the server disconnects the session.

Implementation of

AudioInputProvider.isReady

onAudio()

onAudio(callback): void;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:736

Register a callback to receive audio chunks.

Parameters

Returns

void

Remarks

Only one callback can be registered at a time; subsequent calls replace the previous callback.

Implementation of

AudioInputProvider.onAudio

onSessionStatus()

onSessionStatus(callback): void;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:782

Register a callback for Meet session status updates.

Parameters

Returns

void

Remarks

Invoked whenever the server pushes a sessionStatus resource on the session-control data channel — on admission (STATE_WAITING → STATE_JOINED) and on disconnect (STATE_DISCONNECTED, with a GoogleMeetDisconnectReason). When the session disconnects, the provider automatically stops emission and tears down the WebRTC session.

Example

meet.onSessionStatus((status) => {
  if (status.connectionState === 'STATE_DISCONNECTED') {
    console.log('Meeting over:', status.disconnectReason);
  }
});

pause()

pause(): void;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:700

Temporarily pause audio emission.

Returns

void

Remarks

Used by the turn-taking system during TTS playback. Audio received while paused is silently dropped.

Implementation of

AudioInputProvider.pause

resume()

resume(): void;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:711

Resume audio emission after a pause.

Returns

void

See

pause

Implementation of

AudioInputProvider.resume

start()

start(): void;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:676

Start emitting audio chunks.

Returns

void

Remarks

Meeting audio received while stopped is silently dropped (the WebRTC session keeps flowing; only emission is gated). Must be called after initialize.

Implementation of

AudioInputProvider.start

stop()

stop(): void;

Defined in: src/providers/io/meet/GoogleMeetInput.ts:688

Stop emitting audio chunks.

Returns

void

Remarks

The WebRTC session stays connected — call start to resume emission, or dispose to leave the conference.

Implementation of

AudioInputProvider.stop