tl2012tl/TE_MAN

TE MAN v3.7 – A ComfyUI‑focused AI‑creation suite

What it is

• A plug‑in for ComfyUI, the popular open‑source visual‑programming UI for diffusion models. TE MAN adds a large collection of custom nodes, UI panels and an “asset library” that turn ComfyUI into a full‑featured creative workstation for image‑, video‑ and audio‑generation, prompt engineering, story‑boarding and 3‑D scene planning.
• The package works with local diffusion models and with online LLM/vision APIs (OpenAI, Gemini, Claude, Qwen, MiniMax H3, etc.), letting users switch between on‑device inference and cloud services.

Key capabilities (as of v3.7)

Area	Main features
Canvas & workflow	Unlimited canvas, multi‑task concurrency, auto‑linking of nodes, custom save paths, drag‑and‑drop asset library.
Prompt & script assistance	Prompt‑enhancement nodes for Qwen Image 2.1, MiniMax H3, batch‑prompt management, text‑tool node for 9‑panel storyboard prompts, automatic prompt‑to‑image linking (@ syntax).
Audio	One‑node load‑edit‑save audio (crop, speed, format), audio assets in the library.
Image / video comparison	Side‑by‑side overlay nodes with draggable split line, magnifier, zoom controls; video‑compare node with sync‑play, frame‑sync, speed control and full‑screen mode.
3‑D Director Desk	Interactive 3‑D panel where users place GLB characters, props, and panoramic backgrounds; bone‑mode/OpenPose import; per‑scene JSON saving.
Asset library	Centralised management of images, audio, text prompts; categories, rename, preview, drag back into the canvas.
AI assistant (“构想台”)	Chat‑style panel for brainstorming, script writing, prompt generation; supports multiple API providers, image‑to‑LLM queries, skill‑based workflows (MiniMax H3 video creation, branding, game intro, etc.).
Batch processing	Nodes for batch prompts, batch images, and batch‑enhancement settings (intervals, resume from index, random order).
Compatibility	Requires Python 3.12/3.13, works with ComfyUI v2.6‑plus; optional “nodes 2.0” support (recommended off).

Who would use it

• Artists and video creators who already use ComfyUI for diffusion‑based image generation and want a tighter, “all‑in‑one” UI for story‑boarding, audio editing and 3‑D layout.
• Prompt engineers who need systematic prompt‑enhancement (Qwen, MiniMax H3) and batch execution.
• Small studios that mix local models with cloud APIs and need a single place to manage assets and switch providers.

Installation notes

• The README points to a pre‑built “ComfyUI TE integrated package” hosted on a Chinese file‑share (Quark) and a B‑站 (Bilibili) video for setup.
• Only Python 3.12/3.13 is supported; older Python versions will fail.
• After installing the plug‑in, a new top‑level menu entry “TE MAN” appears in ComfyUI, giving access to the asset library, the “构想台” assistant and the 3‑D director panel.

Limitations / open points

• The project is tightly coupled to the Chinese ecosystem (links to Quark, B‑站, QQ group) and may lack English documentation.
• Some features (e.g., Nodes 2.0) are marked as experimental; the author recommends keeping them disabled.
• No mention of automated testing or CI; stability depends on the author’s updates.

Bottom line TE MAN is a genuine, feature‑rich extension that turns ComfyUI into a multi‑modal creative suite covering prompt engineering, audio/video comparison, batch processing and 3‑D scene planning. It is squarely in the AI‑generated content / generative‑art space and therefore in‑scope.

---

All details above are taken directly from the repository’s README; no external assumptions have been added.