ElevenLabs Pro - Save Audio

Everything the paid nodes in this pack generate lives in memory as an AUDIO dict. At some point you need it to be a file. ElevenLabsPro_SaveAudio is the node that writes any AUDIO clip to disk under ComfyUI/output - wav, mp3, flac, or ogg - with a filename and optional subdirectory you control. It's the pack's output terminal, and the last node in almost every real workflow.

Inputs are minimal: audio, filename_prefix (default elevenlabs), format (wav/mp3/flac/ogg), and subdirectory (optional, nested under ComfyUI/output). The output is saved_path - the absolute path to the file it wrote, which you can feed to a text display so you actually see where your file went. It's marked as an output node, so it shows up as a terminal in the graph.

The mechanism is where the pack's security audit shows its work. The node sanitizes both the subdirectory and filename - stripping anything that isn't alphanumeric/underscore/dash - and explicitly rejects .. traversal or absolute paths. The subdirectory is containment-checked against the output root before anything is written. That's not paranoia: ComfyUI workflows are shared as JSON/PNG, and a malicious workflow could otherwise make the node write wherever it liked. This one can't. It also picks a non-colliding filename with a timestamp and counter, so re-running won't silently overwrite your previous render.

Under the hood it uses soundfile (the libsndfile bindings) to write the waveform - that's one of the pack's two declared dependencies, and it's the one that actually does the heavy lifting here. If you see a "need soundfile / libsndfile installed" error, that's the culprit: pip install soundfile is usually enough, but the native lib sometimes needs a system install on Linux.

Install is the pack standard:

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 install "ComfyUI-ElevenLabs-Pro" via ComfyUI Manager.

Practical notes for using it well: pick wav for maximum compatibility and zero loss (this is the one that plays everywhere), mp3 when you're delivering to a platform that wants compressed files, flac/ogg for archival or web audio. The subdirectory field is how you keep a multi-run session organized - "podcast/ep3" instead of a flat pile of elevenlabs_*.wav. And one thing worth knowing: this node is free and local, so you can re-queue it as much as you like - the cost discipline in this pack comes from the generation nodes, not from saving what they produced. The one trap: it saves what's in the wire at that moment. If you forget to run Normalize or Trim upstream, you're writing exactly what the API gave you - Save Audio won't fix your levels for you.