What alt text is for
Alt text is the text alternative for an image. Screen readers read it aloud, browsers show it when the image fails to load, and search engines use it to understand the image. The test for good alt text is simple: if you replaced the image with the text, would the page still make sense?
- Say what matters here. The same screenshot needs different alt text on a pricing page (“Pricing table with three plans, annual billing selected”) and on a changelog (“New annual billing toggle on the pricing table”).
- Skip “image of”. Screen readers already announce it as an image. Do say “Screenshot of” or “Chart of” when the type matters.
- Include meaningful text. If the image contains words the reader needs, like a button label or a headline, put them in the alt text.
- Leave decorative images empty. Use alt="" so they’re skipped.
How this generator works
Your browser resizes the image to at most 1,280 pixels on the long edge and re-encodes it as a JPEG under 1 MB. We send it, with your optional page topic and keyword, to Meta’s Llama 4 Scout, an open-weight vision model on Cloudflare Workers AI. It writes three options: a literal one, one focused on the image’s purpose, and a short one. We remove “image of” openings, cut anything over 125 characters and drop duplicates. If Llama 4 Scout is unavailable, LLaVA 1.5 describes the image and Llama 3.3 70B writes the options. The image isn’t stored.
Check before you publish
- Look at the image and read each option. Vision models misread small text and sometimes describe things that aren’t there.
- Add names only you know: the product, the feature, the person (with their permission).
- After publishing, run the page through the meta tag checker, and set a proper share image with the Open Graph generator.
The model doesn’t see the rest of your page, so it can’t know what the image is there to show. The page topic field helps; your judgment helps more.