About Us

Not every node in a pack this size is a working piece of a pipeline - some exist purely so the community around the project has somewhere to point. About Us is that node: zero inputs, drop it on the canvas, run it, and it hands back the pack's own self-description as plain text. No LLM call, no external service, nothing to configure.

comfyui_LLM_party is a genuinely large project - the README itself describes it as covering everything from "the most basic LLM multi-tool call" and role setup through GraphRAG knowledge-base management, multi-agent interaction patterns, integrations with QQ/Feishu/Discord, and a full LLM + TTS + ComfyUI production pipeline. Its author, heshengtao (also active under the handle ailm32442), has described the project directly on Reddit as "a set of block-based LLM agent node libraries designed for ComfyUI development" - and has recommended it, unprompted, to people asking for something Dify- or Flowise-like they could run inside ComfyUI rather than as a separate service. That framing is the honest one: this isn't a single-purpose custom node, it's an attempt to make ComfyUI itself a viable place to build LLM agents, with image generation as one thing the agent can reach for rather than the whole point.

What this specific node actually does

Nothing functional. It has no required or optional inputs at all, and its single output is a list of STRING values - meaning it likely returns several lines at once (links, credits, project info) rather than one block of text. Being marked as an output node means it surfaces directly on the canvas when you run it, the same way a debug or display node would, without needing to wire it anywhere else.

Compare it to Mini Story Generator, another zero-friction node in this pack - that one at least tests whether your LLM API credentials actually work, giving you a real "does the install work" signal. About Us doesn't even do that; it's static information, not a connectivity check.

Should you use it?

Honestly, only out of curiosity the first time. Once you know the project's links - its bilibili and YouTube channels, its Discord and QQ group, the author's other project (super-agent-party, a related agent framework with its own RAG/MCP/desktop-companion angle) - there's no reason to keep this node in a real workflow. It costs nothing to run, but it also does nothing for your graph.

How to install it

Search comfyui_LLM_party in ComfyUI Manager and install, then restart. Manual route:

cd ComfyUI/custom_nodes
git clone https://github.com/heshengtao/comfyui_LLM_party.git

Run pip install -r requirements.txt from inside the pack's folder using ComfyUI's own Python, then restart. This node itself needs absolutely nothing beyond the base install - it's static text with no external dependency of its own, in a pack whose other nodes pull in LLM clients, embedding libraries, OCR, and TTS.

Common issues & troubleshooting

There's genuinely not much that can go wrong here - no inputs to misconfigure, no API to fail. If the pack as a whole isn't loading (a not-uncommon report for this project specifically, usually tied to dependency conflicts in the shared requirements.txt across such a large node collection), that's a pack-level install problem rather than anything specific to this node - check that the full install completed cleanly and that ComfyUI's console shows no import errors for comfyui_LLM_party on startup before assuming any individual node, including this one, is broken.