ComfyUI-nkxx

The one node that saves you from pasting keys into every Grsai node

A lazy image input that grows sockets instead of your patience

The payout node — grab the finished Grok video and drop it in your graph

The poll node that tells you if your video job actually happened

One CSV, a pile of Grok video jobs, zero babysitting

Turn a prompt into a Grok video job, then walk away

Debug a monster workflow by running one group, not the whole thing

OpenAI's image model, streaming-tolerant, up to five references

A Gemini chat node with a fixed/random mode that actually means something

Ask Gemini for prompts, get a CSV ready for your batch node

Google's image model on the Grsai channel, 14 reference images and all

A spreadsheet, a concurrency slider, and 50 images later

A whole folder of images, style-shifted with a CSV of prompts

Batch generation that names files after the prompts that made them

Gemini looks at your images and tells you what's there

Press the Midjourney buttons your workflow forgot

Pull the finished MJ grid and auto-split it into four

Midjourney inside ComfyUI, prompt, pad, sref, and all

See every Midjourney task on the Wujiai channel at a glance

Fire a CSV of jobs into a queue that survives restarts

This is what actually hands you the image

The async submit half of the pair

Save images under names you actually choose, not timestamp spam

Recycle a Sora2 video's PID into a character ID — no local files needed

Sora2 text/image-to-video, synchronously, with a progress bar and a download

The node that finally hands you the finished video

'is my video done yet?' — the report card

Submit a job and walk away

Fire 50 Sora2 videos from a spreadsheet without blocking your queue

Turn a snippet of local video into a reusable Sora2 character ID

The spreadsheet batch node for any Wujiai model

Collect the images your background jobs finished

Fire N images at any model and free the queue

Point it at a folder of images + a CSV, get a folder of results back

One node, every Wujiai image model — Gemini, Nano Banana, Grok, GPT, Seedream

Veo 3.1 with first AND last frame control, synchronously

Veo 3.1 with up to three reference images — keep the subject, move the camera

Submit in a second, render in the background

The Wujiai Veo downloader — same fetch node, different channel

The node that finally downloads your clip

Send up to 3 reference images to Veo 3.1 without blocking your queue

A status board for your Wujiai Veo jobs — one string, zero config

The 'is it done yet?' node for your Grsai Veo queue

Veo 3.1 from the Wujiai channel, with three reference images and seven model flavors

Turn a rough idea into a prompt CSV with a cloud LLM — then feed it to the batch nodes

A live catalog of every model this pack can call — no plugin update needed

Grab the next finished Sora 2 video and drop it straight into your graph

Poll your whole Sora 2 batch with one node that never blocks

Fifty Sora 2 videos from one CSV, submitted five at a time

A text or image-to-video job in one node