ElevenLabs Pro - Audio Channels

Half the audio pain in ComfyUI is channel mismatch: TTS comes back mono, your video's soundtrack is stereo, and the moment you try to combine them one side of your head goes quiet. ElevenLabsPro_AudioChannels is the small, local, free fix. It converts any AUDIO dict between mono and stereo, or yanks out a single channel, with zero API calls and zero credits.

Mechanically it's about as pure as a node gets. It takes your audio and a mode dropdown with five options:

• to_mono - averages channels down to one. The default, and the one you'll use most.
• to_stereo - duplicates mono out to both channels (or keeps the first two of a multi-channel file).
• left_only / right_only - keep just that channel.
• swap_channels - flip left and right, for when a sample is recorded backwards from what you expected.

The implementation just reshapes the waveform tensor, computes the mean across channels for mono, or slices/reorders for the rest. Sample rate is preserved, latency is nil, and it needs nothing but torch - which ComfyUI already ships. There's one output, audio, ready to feed the next node.

Why you'd actually reach for it, concretely: Speech to Speech and Audio Isolation both return whatever the API sent back, and a lot of ElevenLabs output is mono. If you then push that into Audio Concat alongside a stereo clip, the concat node handles the mismatch by broadcasting mono up to match - but if you're feeding a stereo consumer directly, or you want a proper stereo bed for a video, run it through to_stereo first. Conversely, if you're about to analyze audio or feed it to a pipeline that expects mono, to_mono saves you a downstream surprise. swap_channels is the classic fix for "my headphones say the music is backwards," which happens more than you'd think with AI-generated samples.

Install is the pack's standard routine:

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

Restart ComfyUI, or use ComfyUI Manager to install "ComfyUI-ElevenLabs-Pro". Requirements are only requests and soundfile, and this node touches neither - it runs entirely on the local waveform tensor.

One thing to keep straight: this is a channel converter, not a resampler and not a loudness tool. If your clips have different sample rates, that's Audio Concat's job (it resamples to match) or you'll want a proper resample step. If your levels are wildly inconsistent, that's Audio Normalize's job. Channel conversion is one narrow, well-defined operation, and this node does exactly that one thing - which is precisely why it's reliable. A small trap with to_stereo: it only expands the first two channels if your input already has more, so don't expect it to rescue a 5.1 file into a proper stereo mix - that's not what it's for.