ACE-Step Show Text

Every ComfyUI pack ends up needing one of these, and ACE-Step's is unapologetically minimal: ACE-Step Show Text takes a STRING, shows it in a read-only text area on the node, and passes the same string through untouched. That's it. No formatting, no file writing, no dependencies beyond the pack itself. It exists because the Text2music Server node's info output is a wall of JSON-ish metadata, and ComfyUI won't display a raw string anywhere without a node that says "show me."

Where it earns its keep is with sample mode. Turn on sample_mode in the Gen Params node, let the LLM decide the bpm, key, duration, time signature, caption, and lyrics - and the only way to see what it actually decided is to look at the server's info text. Wire info → Show Text, run, and read. Same for checking the lyrics the model wrote, or the metadata attached to any generation. Without this node you'd be flying blind on exactly the part of the workflow that's doing the most work.

The inputs and outputs are as small as it gets:

• text - the STRING to display. It's forced to be an input (you can't type into it), so it has to come from a wire.
• text (output) - the same string, passed through. You can tee off this node and keep routing the text elsewhere; it never blocks the flow.

One input, one passthrough output, marked as an output node so it renders in the UI. There's genuinely nothing else to configure.

Install

Same one-liner as the rest of the pack:

cd ComfyUI/custom_nodes
git clone https://github.com/ace-step/ACE-Step-ComfyUI.git
cd ACE-Step-ComfyUI
pip install -r requirements.txt

Restart ComfyUI. (Or: Manager → search ACE-Step-ComfyUI → install.)

Gotchas

There's not much to break. Two things worth knowing: the node only takes wired input, so if you want to display something it has to be connected - there's no free-text field, and that trips people who expect a text box. And since it's a passthrough, the display shows whatever the last run of the graph put through it; if you're not seeing updated info, check that the wire upstream is actually to the Server node's info output and not the audio output. If you need more than this - a markdown renderer, file output, a bigger text view - the ecosystem has fancier Show Text clones, but for checking what ACE-Step decided, this one does the job with zero overhead.