Midjourney ComfyUI
https://github.com/anil-matcha/midjourney-comfyui- Category
- AI Tools
- Rank
- No. 1658Tools index
Previous survey · No. 1642 ·
- Pricing
- Open Source
- Type
- TOOL
- Builder
- Anil-matcha
- GitHub
- 4 stars
- Date
About
ComfyUI custom nodes for Midjourney V7, V8, and Niji via muapi.ai, with 4-image runs and reference image support.
What it does
It turns remote image generation into blocks that fit a visual ComfyUI graph. A block sends prompt controls to MuAPI, waits for completion, downloads the returned images, and converts them into a batch for previewing, saving, upscaling, or further processing.
Why it's ranked here
The focused scope makes the node pack easy to understand and connect. It exposes useful controls, returns both image tensors and request metadata, and handles remote polling internally. However, the documented installation commands reference a different repository, which creates a serious setup risk.
What's good
Every model node presents the same prompt, aspect ratio, style, variation, negative prompt, seed, and reference controls. Users can supply a reference by URL or connect an image directly. Explicit errors cover invalid credentials, exhausted credits, rate limits, failed jobs, and timeouts.
Tradeoffs
Generation depends on MuAPI, an API key, network access, and available service credits. Polling can continue for ten minutes, while failed image downloads become blank placeholders instead of stopping the workflow. The pack covers only three Midjourney endpoints, and the published installation URL appears unrelated to this repository.
How to use it well
Choose it when ComfyUI is your main orchestration surface and remote Midjourney output should feed standard image nodes. Configure credentials once, tune shared controls, then route the batch into previews, savers, or upscalers. Do not choose it for local generation or broad access to unrelated media models.
Technical notes+
README.md documents Manager and manual installation, but both commands use https://github.com/Anil-matcha/Open-Higgsfield-Popcorn rather than the supplied repository URL. requirements.txt declares requests, Pillow, NumPy, and Torch. midjourney_nodes.py submits JSON to model-specific MuAPI endpoints, polls /predictions/{request_id}/result every five seconds for up to 600 seconds, uploads connected reference tensors as JPEG, downloads result URLs into Torch tensors, and zero-pads unequal image dimensions before stacking. Download failures return a 64 by 64 black tensor. __init__.py exports the ComfyUI node mappings, while Midjourney_Example.json connects the credential node, V8 generator, and standard preview node.
Observed
- License
- MIT License
- Primary language
- Python
- Interface
- ComfyUI custom nodes backed by the MuAPI HTTP API
- Install surface
- Clone into the ComfyUI custom nodes directory and install Python requirements
- Runtime requirements
- Python 3.8 or newer, requests, Pillow, NumPy, and Torch
- Authentication
- MuAPI key through a node field, connected credential node, or user configuration
- Included workflow
- Example graph connects credentials to Midjourney V8 and a standard image preview
Read from README.md, requirements.txt, __init__.py, midjourney_nodes.py, LICENSE, Midjourney_Example.json.
What it can do
Generate images using Midjourney V7
Text prompts and parameters → AI-generated images
Generate images using Midjourney V8
Text prompts and parameters → AI-generated images
Generate images using Niji models
Text prompts and parameters → AI-generated images in anime/illustration style
Create 4-image batches
Single text prompt → Set of 4 related generated images
Generate images with reference image guidance
Text prompt and reference image → AI-generated images influenced by reference image
Integrate Midjourney API into ComfyUI workflows
ComfyUI node graph with Midjourney nodes → Executed workflow with generated images
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.