Upfront: I did not build this. It is an open-source project on GitHub (Nutlope/description-generator, 未声明). It is listed here on star count, maintenance activity and how well it fits day-to-day storefront operations.
What it does
- Writes from the image — a vision model reads material, style and use directly from the photo, so you are not filling in attribute fields first.
- Multi-language output — generates several locales in one pass, saving a translation round on multi-language or multi-market stores.
- Short, legible codebase — a Next.js app whose core logic is brief enough to adapt into a WooCommerce bulk script.
- Swappable model — it calls Together AI's vision model by default; pointing it elsewhere is a few lines.
How to use it
Clone it, set the environment variables and run locally:
git clone https://github.com/Nutlope/description-generator
cd description-generator
# create .env with your Together AI key and S3 credentials
# TOGETHER_API_KEY=...
npm install
npm run dev
You need a Together AI API key and an S3 bucket for uploads; .env.example lists everything required. Note the repo declares no licence, so check before commercial use.
When to use it, and when not to
A good starting point for the copywriting step of a bulk-listing pipeline. For a full standard on product imagery and copy, see the E-commerce Visual Copywriting SOP.


