comfyUI-tool-2lab

Only shows the models the 2lab server will accept

Pick a VAE, but only from the platform's shortlist

All five Azure fields, so you can point it at your own deployment

The Azure node that only needs a deployment — because the 2lab server owns the rest

Your own Baidu app credentials, translated text straight into the graph

Same name, no keys — this one rides on the 2lab cloud

Flip one switch, feed a boolean and a string at once

The version with no key field — because it's the 2lab server doing the calling

Plug your Zhipu key in and get a text response for your prompt chain

The boring string-joiner that actually keeps your prompts sane

Cut an image to a region on the canvas, no numeric boxes

Baidu Translate inside the graph — English and Chinese, and that's it

Chinese text in, English out, with your Youdao app key

Azure OpenAI through the Factx marketplace — bring your deployment name

ChatGLM from Zhipu AI, reached through the Factx API marketplace

One float in, three types out — the conversion node you keep reaching for

ImageBatch with nine slots instead of two — batching without the extension pack

Same job as the other scaler, and one footgun of a default

Resize an image to a target size without wrecking its aspect ratio

Force an image to a target edge — the pack's better scaler

An input box that can become an API parameter — the workflow2api idea

Core LoadImage, wearing a workflow2api name tag

One integer in, three types out — and an export flag for API workflows

The seed input that marks itself as user-facing when you publish

A labeled text input that doubles as an exported API parameter

The pack's oddest node — a key input with nowhere to plug in

Slot four snippets into a template and get one string out

One integer, three sockets — the pack's tiny type converter

The latent-side batch node — nine inputs, one tensor, fewer KSampler passes

Load an image from anywhere on disk, not just the input folder

Flip your selection — and yes, core ComfyUI already has this

Call OpenAI from inside your graph — with a model list frozen in late 2023

The pack's save button, tuned for publishing renders

The 'return this' marker for a workflow2api graph

Every generation parameter in one box — the pack's actual centerpiece

Eyeball images straight off the filesystem — no graph detour required

Preview remote images by URL — handy, and worth one security thought

Ship your ComfyUI workflow to the 2lab server as an API

Find-and-replace for your prompts (and any other string that annoys you)

Turn a list of file paths into saved images

Download a whole list of images without leaving the graph

Write text to disk from inside your workflow

Batch-download videos by URL from inside ComfyUI

Finally, a node that tells you your batch size

A preview panel so you can actually see what your string nodes are doing

See the string without a debug print

Load an image from a URL as an actual tensor

A stop button you can wire into your graph

Nine prompts in one box

The simplest text input in the pack

A fixed-text node that just works

Four strings, one list, hello batching

The four-input concat the pack added after the two-input one

Splice lists together for batch pipelines

Kill the stray spaces breaking your prompt

One place for frames and fps

Pixel math for watermark placement, solved

Translate your prompt before it hits the sampler

The Youdao translator that needs no key