audio_mixer

Combine multiple audio inputs into a single mixed output. The output is a continuous audio stream: when every input is idle or no input track is available, audio_mixer emits continuous silence so downstream encoders and muxers keep a stable timeline. Idle channels use Zero Gain Optimization to skip unnecessary downstream DSP.

Example

{
  "name": "mix_audio",
  "type": "audio_mixer",
  "inputs": ["room_src"],
  "config": {
    "sampleRate": 48000,
    "channels": 2,
    "audioLevels": { "enabled": true, "intervalMs": 1000 }
  }
}

Config

Field	Type	Default	Description
sampleRate	int	48000	Output sample rate. Values outside the allowed set are snapped to the nearest of {8000, 16000, 24000, 32000, 44100, 48000}.
channels	int	2	1 or 2 (clamped).
audioLevels	object	off	Optional per-identity volume reporting for clients. Omitted = disabled.

audioLevels

Field	Type	Default	Description
enabled	bool	false	Emit quantized volume tiers to supporting sinks (RTC data channel or in-band video metadata for media sinks).
intervalMs	int	1000	Report interval in ms. Clamped to 200–2000.
levels	int	5	Number of discrete tiers; output is 0…levels-1 (0 = silence). Clamped to 2–10.
floorDb	int	-50	Silence threshold in dBFS, not an output volume. Values are usually negative; positive values are normalized to negative (e.g. 50 → -50). Less-negative values such as -30 require louder input than more-negative values such as -70.

Pricing

$0.0005/min of actual processing time. Node pricing.