orion4d/orion4d_metaprompt — explained in plain English
Analysis updated 2026-05-18
Build complex image-generation prompts from reusable text lists without cluttering the ComfyUI canvas.
Automatically expand a short prompt idea into a detailed one using a local language model.
Generate a text prompt from an existing image using a local vision model.
Organize, clean, and export large batches of prompt fragments with the standalone List Constructor tool.
| orion4d/orion4d_metaprompt | agi-ruby/ai-gpt_image2-seedance_2.0-video-skills | arindam200/wisprgemma | |
|---|---|---|---|
| Stars | 10 | 10 | 10 |
| Language | JavaScript | JavaScript | JavaScript |
| Setup difficulty | moderate | easy | moderate |
| Complexity | 3/5 | 1/5 | 3/5 |
| Audience | vibe coder | vibe coder | general |
Figures from each repo's GitHub metadata at analysis time.
Requires a running local Ollama instance with the desired text or vision models already pulled.
Orion4D MetaPrompt is a set of custom add-on tools for ComfyUI, a visual node-based interface used to build AI image generation workflows. Its purpose is to make writing and organizing prompts easier by adding pop-up windows directly inside ComfyUI, instead of forcing you to wire together dozens of small text nodes on the canvas. The main piece is the MetaPrompt node, which builds a prompt out of text lists loaded from txt or csv files along with manual text chunks you type in. It opens in a full-screen pop-up where you can add and reorder as many text blocks as you like, and for each list you can choose whether it always uses the same line, picks a random line every run, or steps through lines one at a time in order. A second node, MetaPrompt Ollama, builds on the same block system but adds a step that sends the assembled prompt to a locally running language model through Ollama, which expands a short idea into a more detailed prompt with added lighting, camera, and style description. If the local model gives back an empty answer or just repeats the input, it automatically retries or falls back to the original text so the workflow does not break. A third node, ImageToPrompt Ollama, works the other way around: it takes an image, either connected from earlier in the workflow or read from a folder of image files, and asks a local vision-capable model to describe it in words, producing a usable prompt from a picture. Alongside these ComfyUI nodes, the project includes a separate browser based tool called List Constructor for preparing prompt lists before using them. It lets you type or paste prompt fragments, label and color-code them, remove duplicates and empty entries, search and replace text, sort entries, and export the finished list as a txt or csv file, or save your whole session as a reusable configuration file. All of the nodes rely on you having Ollama installed and running locally with the language or vision models you want already downloaded. File paths used by the nodes are checked against a safety function before anything is read or written, to prevent them from reaching outside the folders you have configured. No license is stated in the project.
Custom ComfyUI nodes that build and organize image-generation prompts, including local AI prompt expansion and image-to-prompt captioning via Ollama.
Mainly JavaScript. The stack also includes JavaScript, ComfyUI, Ollama.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly vibe coder.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.