Question 1

Are my images uploaded to a server?

Accepted Answer

No. This AI image captioner runs entirely in your browser using WebAssembly. Your images are processed locally via the ViT-GPT2 model and never leave your device.

Question 2

What is alt-text and why is it important?

Accepted Answer

Alt-text (alternative text) describes an image for users who cannot see it — including screen reader users. WCAG 2.1 requires alt-text for all informative images. This tool generates <img alt="..."> HTML snippets for direct use.

Question 3

How accurate are the captions?

Accepted Answer

ViT-GPT2 is trained on the COCO dataset and achieves CIDEr scores comparable to state-of-the-art models. It works best with well-lit, clear photographs. Abstract art, text-heavy images and diagrams may produce less accurate results.

Question 4

Why is the model so large (~250MB)?

Accepted Answer

Vision-language models must encode images (ViT encoder) and generate text (GPT-2 decoder). The combined model with quantization is ~250MB. We recommend WiFi for first-time download. After caching, captions take ~1-3 seconds each.

Metric	Value
Model size (INT8)	~250MB
CIDEr score	103.2 (COCO test)
Avg generation time	1.8s / image
Model load (cold start)	~45s (CDN) / <2s (cached)

AI Image Captioner

Upload Images

How It Works

Frequently Asked Questions

Are my images uploaded?

What is alt-text?

Our Test Data

When Not To Use This Tool