Runware Audio Settings

Audio nodes in this pack - the text-to-audio and music models under Runware/Audio - accept an audioSettings socket, and this builder is what fills it. It's the smallest builder in the pack: three settings, two of which you'll probably never touch. Its whole job is to produce a RUNWARE_AUDIOSETTINGS value and hand it to an audio model node so your generated audio comes out the way you want instead of whatever the model defaults to.

The settings

• channels - the one you'll actually change. An enum of 1 (mono) or 2 (stereo), default stereo. If you're generating a voiceover or a podcast clip, mono is honest and halves the file; if it's music or anything with width, leave stereo. This is the only knob with no gate - it's always sent.
• bitrate (+ bitrate_value, in kbps) - off by default, which means "use the model's default." Flip the toggle and you can pin it, say 8 kbps for a tiny voice note or higher for music. The tooltip floor is 8 kbps.
• sampleRate (+ sampleRate_value, in Hz) - same gate pattern. Off uses the model's default; on lets you request a specific rate. Most people never need this - the model's trained output rate is the right one.

The gate pattern, once, because it repeats everywhere

See that bitrate boolean with a bitrate_value next to it? That's the pack's off-by-default convention, used across every builder: the toggle gates the paired value, and "off" means "don't send the field, let the model decide." It exists so a default workflow never sends a half-set parameter that breaks a validation rule. It's not a bug, it's a design - and once you've met it here, you'll stop being confused by it on the accelerator, outpaint, and PuLID builders.

Output and wiring

Output is RUNWARE_AUDIOSETTINGS. Wire it into the audioSettings socket of any audio model node that exposes one - about 21 model nodes in the pack do. If the audio model you're using doesn't have that socket, you don't need this node.

Install

Same one-time pack install as everything else here:

cd ComfyUI/custom_nodes
git clone https://github.com/Runware/ComfyUI-Runware
pip install -r ComfyUI-Runware/requirements.txt

Restart, add your Runware API key (Settings → Runware API key, or RUNWARE_API_KEY).

The honest take

This is a thin, quiet node, and that's fine - most of the time the honest answer is "wire it in, set channels to mono if you're doing voice, and leave the rest." The bitrate and sampleRate toggles are there for the cases where the model's default isn't what your pipeline needs, and that's exactly the right scope. Don't feel like you're underusing it by leaving toggles off; in this pack, "off" is the intended default.