Guide
Alt text, caption and long description: three jobs
Alt text, a caption and a long description all describe the same picture, and that is where the resemblance ends. They are read by different people, at different moments, for different reasons, and writing one as if it were another is the most common mistake on the page.
Who reads which
The alt attribute is read instead of the picture. Someone using a screen reader reaches the image and hears the alt text where a sighted reader would glance at the photograph. It is short because it is a substitute, not an addition: one sentence that does the job the picture was doing in that spot.
The caption is read by everyone, in addition to the picture. It is visible text under or beside the image, so a sighted reader sees both and a screen reader user hears both. Its job is what the picture cannot do on its own — the name of the place, the date, who took it, why it is here.
The long description is read by someone who has chosen to read it. It is for the picture whose content is too structured for one sentence: a chart, a diagram, a map. The alt says what the thing is; the long description gives the reading, in the order a listener needs.
One photograph, three texts
The alt text
“A fisherman pulling a net into a wooden boat at dawn.” What is in the frame, what is happening, where — in that order, and then stop. No “image of”, because the screen reader has already said it is an image.
The caption
“Kovalam, February 2026: the sardine catch has halved in five years, fishermen say.” Nothing in the caption repeats the alt. It adds place, date and the reason the photograph is in the article.
The long description
Not needed for this photograph. If the article also printed a chart of the catch by year, that chart would get a short alt (“Line chart of the sardine catch in Kerala, 2021–2026”) and a long description beside it: what the chart shows overall, then the values in order.