Vision-language · 100% in your browser · no server

In-Browser Image Captioning

Upload a photo and a vision-language model describes it in natural language — running entirely on your device via Transformers.js. Mirrors the VLM / document-understanding work I ship in production. Built by Mohammed Haitham El Zamzamy.

or try an example ↓

Model loads on first run (~once, then cached).

AI caption

🤗 Hugging Face Static Space · Transformers.js · model Xenova/vit-gpt2-image-captioning
Portfolio · GitHub · LinkedIn