Nodes/ComfyUI-Pollinations/Play Audio 🔊
ComfyUI Node

Play Audio 🔊

Hear Your TTS Without Leaving ComfyUI

By 1038lab·Created about a year ago·Updated about a year ago· 48
Play Audio 🔊
  • AUDIO
  • AUDIO
  • AUDIO_PATH
audio_path
autoplaytrue

Play Audio 🔊 is the smallest node in the ComfyUI-Pollinations pack, and you'll probably only ever meet it on purpose once - the moment your first Premium Text-to-Speech clip comes back and you realize ComfyUI doesn't really have a built-in "play this" button. This node is that button. It takes an audio signal (or a file path) and plays it on the machine actually running ComfyUI.

What it does

The node is dead simple. It accepts an AUDIO signal - the same kind the premium TTS node outputs - or an audio_path pointing at a sound file on disk. If you hand it a waveform, it converts it through pydub into a playable segment; if you give it a path, it loads the file directly. Then, if autoplay is on (it is, by default), it plays the result immediately when the workflow runs.

Its outputs just pass the buck: AUDIO hands the segment through for chaining into other audio nodes, and AUDIO_PATH tells you the file it played (empty string when it came from a live waveform instead).

The gotcha that bites everyone

Here's where the wheels come off for a lot of people. The pack's requirements.txt only lists requests and pydub. It does not install anything that can actually make noise. On Windows the node cheats via the built-in winsound module, so it just works. On Linux and macOS, pydub's playback needs a backend - simpleaudio, pyaudio, or ffplay on your PATH - and none of them ship with the pack. If you're on a non-Windows box and hear nothing, that's why. Fix it with:

/ComfyUI/python_embeded/python -m pip install simpleaudio

Restart ComfyUI after installing. And if you're running on a headless server with no audio device, no amount of installing will help - there's no speaker to reach. Playback is host-side, not in your browser.

How to use it

Drop the audio output of a TTS or audio node into the AUDIO input, leave autoplay on, and run. The node is marked as an output node that always executes, so it won't silently skip itself. Keep autoplay on if you're testing voices; flip it off if you're building a workflow you'd rather not deafen yourself with on every run.

That's the whole thing. It's a utility, not a star - but it's the utility that makes the pack's TTS feature actually usable, and it's the first thing you'll reach for when the silence is suspicious.

Category🧪AILab/🌸Pollinations

Inputs (3)

NameTypeDefaultDescription
AUDIOoptAUDIO
audio_pathoptSTRING
autoplayoptBOOLEANtrue

Outputs (2)

NameTypeDescription
AUDIOAUDIO
AUDIO_PATHSTRING