Captions, alt text and figcaption: which does what

A caption and alt text do different jobs for different readers. A caption is visible content that every reader gets, and it adds context: who is in the photo, where it was taken, who took it, or what a chart means. Alt text is a replacement that stands in for the image when someone cannot see it, such as a screen reader user, or anyone whose image has not loaded. Because alt text replaces the image, the HTML Standard says it should not contain text that could be considered the image's caption, and it should not repeat the prose beside the image. A figure with a figcaption does not change that: WebAIM says the image inside it still needs alt text of its own.

Getting this right matters because missing alt text is common. The February 2026 WebAIM Million found missing alternative text on 53.1% of home pages.

What is a caption?

A caption is visible text attached to an image that tells every reader what the picture is or why it is on the page. It is ordinary page content, not something hidden for a subset of readers. Typical captions give a name, a place, a date, a credit, the medium of an artwork or the takeaway of a chart. The W3C's groups of images tutorial, for example, captions its castle pictures with the medium, the artist and the year.

A caption is optional, and alt text is not. Add a caption where a reader benefits from context, not to every image.

Video uses the word differently. In WCAG, captions are synchronized text for the speech and the non-speech sounds needed to understand media, and success criterion 1.2.2 (Level A) requires them for prerecorded audio in synchronized media. An image caption does not meet that requirement, and a video caption does nothing for a photo. Our guide to video captions and transcripts covers the media side.

What is alt text?

Alt text is the text in an image's alt attribute that serves as a replacement for the picture. WebAIM lists what it does: screen readers announce it in place of the image, the browser shows it when an image fails to load or is blocked, and search engines use it. WebAIM also notes that a text alternative can sit in visible text near the image, and that every image should still carry an alt attribute, even an empty one.

The W3C's tips for writing alt text give a simple test: imagine reading the page aloud over the phone to someone who needs to understand it. Describe the image's purpose on this page as concisely as possible, with the most important information at the beginning. If that takes more than a short phrase or sentence, use a long description method instead. Usually skip words such as "image", "icon" or "picture", because screen readers already announce the image. For the length question, see how long alt text should be, and for the full method see our alt text guide.

What is figcaption?

figcaption is an HTML element, not a kind of text. It holds a caption inside a figure. The HTML Standard defines a figure as self-contained content, optionally with a caption, that is typically referenced as a single unit from the main flow of the document, and says the figcaption must be the first or last child of the figure. It also says an image that is part of a sentence does not belong in a figure.

The figcaption names the figure, not the image. MDN says the caption provides the accessible name of the figure, and WebAIM says the image inside still needs its own alt text. So figcaption is where a caption goes in the markup, and it does not stand in for alt when you can describe the image. The HTML Standard also lists the title attribute as another way to supply a caption, but notes that many user agents do not expose it in an accessible way, for example by needing a mouse to show the tooltip. We use a visible figcaption instead.

Caption, alt text and figcaption side by side

How caption and alt text work together

Give each one a different job. The alt text says what the image shows and does on this page. The caption says what the page wants to add: names, credit, date, place or the point of the picture. The W3C groups tutorial shows the split. Its alt text for the oldest castle picture describes what is drawn, a single tower inside a tall wall, while the caption gives the medium, the artist and the year. Neither could stand in for the other.

If the alt text repeats the caption, anyone who gets the alt text in place of the image hears the same words twice. The HTML Standard picks out this exact pattern as a bad use of alt text: it duplicates the caption for users who do not get the image and tells them nothing more. Here is the repeat and a fix, using an invented photo of volunteers:

<!-- Repeats itself -->
<figure>
  <img src="planting.jpg"
       alt="Volunteers planting oak saplings along the river path">
  <figcaption>Volunteers planting oak saplings along the river path.</figcaption>
</figure>

<!-- Each says something different -->
<figure>
  <img src="planting.jpg"
       alt="People in rain jackets kneel in a line on a gravel path, pressing soil around young trees.">
  <figcaption>Volunteers planting oak saplings along the river path. Photo: the garden club.</figcaption>
</figure>

In the second version the caption carries the context and the credit, and the alt text carries the picture. If the alt text adds nothing to the caption, it is a repeat.

Which text goes where

The sources lean slightly differently on the second row. The HTML Standard says an image that only reinforces nearby text takes an empty alt, but that an image given a caption in a figure needs alt text, and that an image whose caption gives its title must not have an empty alt. The W3C's informative images tutorial likewise says an image might be considered redundant, and so decorative, when the text already explains what it shows. Our rule of thumb: in a figure with a caption, write alt text that adds to the caption, and keep alt="" for images that merely echo the prose beside them and have no caption of their own. For more cases, see alt text examples.

The W3C's groups tutorial nests figures, one per picture plus one for the set, and adds role="group" to each, while warning that support for that attribute and value varies. ARIA in HTML allows no role other than figure on a figure that contains a figcaption, so we leave the role off. The nesting and the captions do the work.

Mistakes to avoid

• Copying the caption into alt. It makes the reader hear it twice. Describe the picture instead, and use an empty alt only when the image merely echoes the prose beside it.
• Describing the picture only in a caption. A caption reaches every reader, but a reader who cannot see the picture still needs to know what the caption refers to. The HTML Standard says alt text spares that reader the confusion. Put what the picture shows in the alt text, and let the caption add to it.
• Starting alt text with "image of" or "picture of". The W3C says there is usually no need for words like image, icon or picture, because screen readers announce the image. Naming the medium is fine when it matters, such as a painting versus a photograph.
• Leaving the text in the picture. A caption drawn into the image's pixels is not text. Put real text beside the image, and see images of text.
• Leaving alt off because a caption exists. W3C's failure technique F65 covers omitting the alt attribute or any other text alternative on an image. The HTML Standard allows an omitted alt only in narrow cases, such as an image whose contents the author cannot know.

Related image types have their own rules: screenshots, icons, logos and social media posts.

What WCAG asks for

Success criterion 1.1.1 Non-text Content (Level A) requires that non-text content has a text alternative that serves the equivalent purpose. W3C's Understanding page separates images where a short description can do the job from images, such as charts, where it cannot. A caption can be part of the answer for the second kind. W3C's technique G74 describes a caption under a chart with the long description in the following paragraph, and the short alt text tells the reader where that description is.

This guide covers the writing, not the law. For the laws that point to WCAG, see our overview of accessibility laws worldwide, and for the full list of criteria see the WCAG 2.2 checklist. None of this is legal advice.

Do captions and alt text help SEO?

Both are read by search engines, and neither is a ranking promise. Google's image guidance says it extracts information about an image's subject from the page content, including captions and image titles, and that alt text is the most important attribute for providing more metadata about an image. Google uses alt text together with computer vision and the page's contents, and it warns against stuffing alt attributes with keywords. The SEO advice and the accessibility advice point the same way: accurate alt text, plus a caption that says something true. See alt text and SEO for the longer version.

Check it yourself

• Run the alt text checker on a page. It downloads the page's images and judges whether each alt text is true of its picture, for the first six described images. Whether the alt text and the caption read well together is an editorial call, so read them aloud in order afterward.
• If an image has no alt text yet, the alt text generator writes a draft for an uploaded image. Check the draft against your caption, and read how accurate AI alt text is first.
• For a first look at a whole page, Scan my site for free.

Keep it checked

A page that is right today can drift after the next upload. GotAlt checks the pages you monitor every week. It checks alt text against the image itself, for up to six images per page, runs its rule checks, opens monitored pages in a browser for contrast and other render-dependent checks (see what we check), and reports what changed. Plans start from $69 a month: Audit my whole site from $69/mo.