Analysis updated 2026-05-18
Generate and iteratively refine marketing images inside Codex by drawing annotations directly on the canvas.
Create Xiaohongshu covers, YouTube thumbnails, or Amazon product image sets using built-in workflow skills.
Resize one image to all required dimensions for multiple social media platforms in a single step.
Use the canvas as a visual annotation board inside Codex to iterate on design concepts without leaving your AI tool.
| binghe1980/ai-canvas | suyancc/openai-plus-vxt | germondai/trawl | |
|---|---|---|---|
| Stars | 182 | 217 | 218 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 2/5 | 3/5 | 3/5 |
| Audience | designer | general | ops devops |
Figures from each repo's GitHub metadata at analysis time.
Requires Codex (the AI coding assistant) and Node.js 20+, installed with two commands then restart Codex.
AI Canvas is a plugin for Codex, an AI coding assistant, that adds an infinite drawing board where you can generate images with AI, draw annotations on them to request changes, and use preset business workflows to produce marketing materials. The whole thing runs locally on your machine. The basic loop works like this: you tell Codex to open the canvas, it generates an image from your description, and then you mark up that image directly on the canvas by drawing arrows, circles, or writing notes, then click a button to send those annotations back to Codex for revision. Each new version appears next to the old one so you can compare them side by side. There are six built-in Skills, which are structured workflows for specific business tasks. These include creating Xiaohongshu (a Chinese social media platform) cover images, YouTube thumbnails, product marketing image sets for Amazon and Shopify, logo and brand identity concepts, marketing brochures, and a cross-platform resize tool that takes one image and reformats it for multiple social media dimensions at once. Setup involves two commands to install the plugin from GitHub into Codex, then restarting Codex. The canvas service runs on your own computer at a local address, so no data is sent to any hosted backend by default. Generated images and canvas state are saved in a hidden folder in your project directory. The README is primarily in Chinese with an English version linked. The license is MIT, meaning you can use and modify it freely for any purpose.
AI Canvas is a Codex plugin that adds an infinite canvas for AI image generation, annotation-based editing, and preset business workflows like social media covers and product image sets.
Mainly TypeScript. The stack also includes TypeScript, Node.js, pnpm.
Use freely for any purpose, including commercial, as long as you keep the copyright notice (MIT).
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly designer.
This repo across BitVibe Labs
Verify against the repo before relying on details.