Mureka AI Generate Lyrics (Official Edition)

This is the boring node in the Mureka set, and I mean that as a compliment. Every other Mureka node in BizyAirPlus spits out audio; this one just writes lyrics. It's an LLM call through the Mureka API - give it a description of the song you want, get back lyric text you can hand to the actual generation nodes. In a music pipeline, a dedicated lyricist node is genuinely more useful than making one of your general-purpose LLM nodes do double duty.

What it is

Mureka is an AI music generation platform in the Suno/Udio lane, and these "Official Edition" nodes are its official API surfaced inside ComfyUI via BizyAirPlus. This one is the text brain: prompt in, result out. No audio, no media upload, no metered render time in the same way - just a language-model pass.

How it works

You describe what you want in natural language - the mood, the theme, the structure ("a melancholic verse-chorus pop song about a red-eye flight, 90s alt-rock vibe") - and Mureka's LLM returns complete lyrics, typically with verse/chorus scaffolding. The node is an output node, so it works standalone or wired into whatever's downstream: you can copy the text, or route it into a Mureka text-to-audio node's lyrics input to actually sing it.

The inputs and outputs

It's about as thin as a node gets:

• prompt - required, multiline. The richer your description, the more usable the lyrics.
• outputs - result (the lyric text as a STRING) and request_id.

api_config and skip_error are the pack standard. That's the whole thing.

Installing

Same as every BizyAirPlus node:

cd ComfyUI/custom_nodes
git clone https://github.com/siliconflow/BizyAirPlus.git
cd BizyAirPlus
python -m pip install -r requirements.txt

Restart ComfyUI, turn the BizyAirPlus button ON, and enter your API key from bizyair.ai (stored at ~/.BizyAirPlus/apikey.ini).

Where people get burned

It's a remote LLM, so treat the output as a draft - the "lyrics" will be structurally correct and sometimes genuinely good, sometimes boilerplate. The Mureka platform enforces content filtering on its music generation, so don't waste a request on lyrics you know the downstream nodes will refuse to sing. And because there's no audio here, remember this node doesn't validate anything about singability - the generation nodes are where you find out.