WanVideo Lora List → String

Wan 2.2 is a two-expert model, and that split shows up everywhere in the tooling. The high-noise expert handles motion and composition, the low-noise expert handles detail - and they carry separate LoRA stacks. It's a distinction with consequences: the community consensus is that speed LoRAs go on the low-noise pass at 0.4–0.6 strength and not on the high-noise pass, or they chew up everything good about the generation. When you're running that setup, knowing exactly which LoRA sat on which expert is the difference between "reproduce this" and "guess."

WanVideoLoraListToString takes those per-expert LoRA lists and turns them into readable strings. You feed it the high_loras and low_loras sockets from Kijai's WanVideoWrapper (both WANVIDLORA type, both optional), pick a format, and it returns a pair of STRING outputs - high_loras and low_loras - with the names and strengths formatted for a human.

How it formats

Each LoRA entry is rendered as name(strength), strength to two decimal places. The format dropdown changes the separator:

• list - one LoRA per line (the default; good for a report's "LoRA:" section).
• comma_separated - one line, compact.
• numbered - 1. name(0.5) style, useful if order matters.

A connected-but-empty side returns an empty string rather than an error, and an unwired side returns empty too, so you can hook up both expert stacks and only fill in what you use.

Why it exists

Same story as the pack's model helper: the report node can already digest WANVIDLORA objects internally through its high_loras_obj / low_loras_obj inputs. This node exists for the cases where you want those strings somewhere the object won't fit - a text note, a filename, a generic metadata field, or a report node from a different pack. It's the "I need the LoRA list as prose" adapter.

Install and wiring

It's part of blkot/ComfyUI-tgBridge. ComfyUI Manager: search "tgBridge", or:

cd ComfyUI/custom_nodes
git clone https://github.com/blkot/ComfyUI-tgBridge
cd ComfyUI-tgBridge
pip install -r requirements.txt

Restart ComfyUI; it shows up under the Telegram category, and it's not an output node, so it sits happily mid-graph. No models to download, just the pack's two light dependencies (Pillow, requests).

Fair warning for a utility node: it's a formatting convenience, not a workflow essential. If you're already feeding the *_obj inputs into Telegram Workflow Report, this node is redundant. But if you've ever stared at a Wan render and couldn't recall whether the speed LoRA crept onto the high-noise pass, having the high/low split as plain text in your report is worth the one extra wire.