cocktailpeanut/ideoprompt — explained in plain English
Analysis updated 2026-05-18
Convert a plain description like 'a sunset over mountains' into a valid JSON prompt for Ideogram 4.
Drag and resize element bounding boxes on a visual grid before generating an image.
Call the tool's HTTP API from curl, JavaScript, or Python to automate prompt generation.
| cocktailpeanut/ideoprompt | abdulrdeveloper/chai-aur-react | charlesdove977/advertising-ops | |
|---|---|---|---|
| Stars | 14 | 14 | 14 |
| Language | JavaScript | JavaScript | JavaScript |
| Setup difficulty | easy | easy | moderate |
| Complexity | 2/5 | 1/5 | 3/5 |
| Audience | designer | vibe coder | pm founder |
Figures from each repo's GitHub metadata at analysis time.
One-click Pinokio install handles Node.js and the ~2.5GB model download automatically.
Ideoprompt is a tool that converts a plain-English image description into a structured JSON prompt for the Ideogram 4 image generator. Ideogram 4 works better with JSON-formatted prompts than plain text, but the JSON format has strict rules that are easy to get wrong by hand. This tool automates the conversion and guarantees a valid result every time. The app runs entirely on your own machine. A small language model (Qwen3-4B) handles the conversion, and it runs via llama.cpp bindings bundled into the JavaScript app, so no internet connection or external API is needed after the initial model download. The guarantee of valid output comes from three layers working together. First, the model's token sampling is constrained by a grammar at generation time, so it cannot produce JSON with wrong keys, wrong order, or wrong types. Second, the output is normalized deterministically: hex colors are uppercased, bounding box coordinates are clamped to the correct range, and palette sizes are capped. Third, the result must pass a schema validation check before it is shown to you, if it fails, the app tries again once with the error details in context. The web interface lets you type a description, click Generate, and see a layout preview showing each element's bounding box on a grid. You can drag boxes to reposition them or resize them, and the JSON updates live to match your adjustments. When you are happy with the layout, you copy the JSON and paste it directly into Ideogram 4. Installation is either one click via Pinokio (which handles Node.js, dependencies, and the model download automatically) or manually via a few terminal commands. The model file is about 2.5 GB. The app also exposes a small HTTP API for programmatic use from curl, JavaScript, or Python.
Ideoprompt turns a plain-English image description into a valid, structured JSON prompt for the Ideogram 4 image generator, running fully offline on your own machine.
Mainly JavaScript. The stack also includes JavaScript, llama.cpp, Qwen3-4B.
Setup difficulty is rated easy, with roughly 30min to a first successful run.
Mainly designer.
This repo across BitVibe Labs
Verify against the repo before relying on details.