Digital Products and Accessibility 5 August 2026
Alt Text Localisation: What Does a User Who Cannot See the Image Read?
The alt attribute behind an image is the only text a screen reader speaks aloud. If it is skipped in translation, the user never finds out what the image was.
Alt text localisation means handling, within the translation process, the short description embedded alongside an image that gives its spoken equivalent to a user who cannot see it. This text is invisible on screen; it sits inside the HTML code, in the alt attribute of the img tag. For a visually impaired user working with a screen reader, it is the only thing that reveals the image is there at all.
Contents
The problem is that, because this text cannot be seen, it is easily skipped in translation. When an image is localised, when the wording on it changes or when it is replaced altogether, the alt text is most often left in the source language. Yet this small description decides, for one group of users, whether the content exists at all; skipping it means that for that group a section of the page is simply not there. In this article we look at where alt text is embedded, which standard governs it, the mistakes made most often and where it sits in the process.
Where is alt text embedded?
Digital products have two layers: the layer visible on screen (interface, text, images) and the invisible layer (HTML code, attributes). Many accessibility elements are hidden in that invisible layer, and alt text is the most classic example of them. In the example given in the source book, the code for a logo reads as follows: inside the image tag it says alt="Routledge Home". The text to be translated is the phrase between the quotation marks.
Text alternatives of this kind can easily be extracted by translation memory and localisation tools and put in front of the translator. But when the tool presents one as a string, the translator does not see the image. This is where the loss of context begins: you have only a few words in hand and you have to work out which image they describe.
This split between the visible and the invisible layer also explains why alt text goes unnoticed. An editor checking the page by eye does not spot it, because nothing on screen points to it. Alt text is therefore an element seen only by someone who opens the code or an automated checker, and for exactly that reason it has to be brought into the translation scope separately.
WCAG 1.1.1 and non-text content
WCAG, the most widely used framework for web accessibility, is built around four principles: perceivable, operable, understandable and robust. The first criterion under the first principle, 1.1.1 Non-text Content, requires every non-text element presented to the user to have an equivalent text alternative. Images, charts, icon buttons and tables all fall within its scope.
This criterion requires the alt text to serve the same function as the image. The aim, in other words, is not to describe the image word for word but to render its communicative role on the page in text. The alt text of a warning icon is the warning it carries, not its colour or shape. A translator who misses this distinction produces text that is technically correct but functionally empty.
A similar criterion covers time-based media. For prerecorded audio-only or video-only content, a transcript conveying equivalent information is expected. That way a user with a visual or hearing impairment can reach the information the media carries through text instead. Alt text is the most visible member of this wider family of text alternatives.
Common mistakes in alt text localisation
WCAG explicitly lists practices that do not meet the criterion as failure techniques. The best-known mistake is to use a generic placeholder instead of alt text: writing just image or video. Another classic is to leave the file name as the text alternative; a screen reader announcing presentation_video1 to the user conveys nothing.
| Weak alt text | Functional alt text |
|---|---|
| image | The exterior of the Kızılay office |
| img_2043.jpg | Example of a sworn translation stamp |
| click | Open quote form button |
As the table shows, the difference lies in whether the user can picture the image or not. Research has shown that translated pages often end up less accessible than their original versions; turning that picture around is in the hands of localisation. A well-written alt text makes a page not merely compliant but genuinely usable.
Another subtle mistake is making alt text longer than it needs to be. Because the screen reader speaks every single word, unnecessary detail wears the user out. Good alt text is short, concise and focused on the function of the image; it neither leaves anything out nor drowns the user in needless description.
When the image changes, the alt text changes too
During localisation, images are frequently adapted or replaced. A screenshot carrying text in the source language is retaken for the target language; a photograph suited to the local user is put in its place. The alt text then has to match the new image. Otherwise the screen reader describes an image that is no longer on the page.
This is the point at which localisation calls for a technical and cultural intervention rather than a linguistic one. Screenshots on help pages are expected to be exactly consistent with the translation in the interface. When that internal consistency between image and text breaks down, the user notices that the screen in front of them does not match the description they are reading, and their confidence is shaken.
Alt text should therefore be handled together with the image decisions, not at the end of the translation project. Whether an image will be replaced, what text it carries and what its local equivalent will be should all be discussed around the same table. Alt text translated in isolation from the image runs the risk of describing the wrong picture, however well it is written.
When is a decorative image left empty?
Not every image needs alt text. Elements that are purely decorative or there for layout, such as the thin rules that separate sections, carry no information. For these, WCAG recommends leaving the alt attribute empty so that the screen reader skips the element. Otherwise the user has to listen again and again to descriptions that mean nothing.
This decision is easily missed in localisation. A translator who assumes that the string in front of them must at all costs be translated may add an unnecessary description to a decorative element. The right response is to ask whether that element carries information. Making content more accessible sometimes means choosing to write nothing.
Who writes alt text, and where does it sit in the process?
The best author of alt text is the person who can see the image in place. The ideal workflow is therefore one in which the translator works in an environment that shows the images. If the source product has no alt text at all, localisation should treat this not as a shortfall but as an opportunity: the translated product can be made more accessible than the original.
Within the process, alt text should be addressed at both the translation and the quality control stage. Automated accessibility checkers guide the translator by listing the elements whose alt text is missing or faulty. This check can be repeated before and after delivery. A gap in the invisible layer is caught that way before it ever reaches the user.
Alt text for complex images
Not all images can be described at the same level. A one-sentence alt text is enough for a simple photograph, but for a chart, an infographic or a flow diagram a short sentence cannot carry the information. For images of this kind the alt text should summarise what the image says, while the detailed data should be given in the body text or in a separate description close by.
This distinction is easily lost in localisation. If a chart that has a long description in the source product is left with only a short alt text in the target language, the visually impaired user reaches no more than the surface of the data. With complex images the translator therefore has to handle both layers: the short, summarising alt text and the long description that conveys all the information in the image.
This also shows that alt text is not only a translation task but a content decision. Deciding which of the information in an image goes into the text and which counts as decorative calls for judgement. Made well, that judgement can leave the translated content even more accessible than the original; made badly, the most important data disappears in silence.
The Türkiye context
In Türkiye, accessibility is defined as a general principle by Law No. 5378 on Persons with Disabilities, which provides for public services to be accessible. In web accessibility, the international WCAG principles and the accessibility standards in Europe are increasingly taken as references in public sector and corporate projects. Invisible elements such as alt text, however, are neglected, because checks are usually carried out by eye.
The structure of Turkish affects the work too. Because it is an agglutinative language, a concept that takes two words in English can turn into a single long word in Turkish, or the other way round. Alt text should therefore give the equivalent the image has for a Turkish-speaking user rather than follow the source string one to one. On top of that, images of local documents such as stamps, seals and apostilles are meaningless to the user and pass on incomplete information when they are not described with the right term.
What does this mean for you?
If you are having your site or your app translated, translating the visible text is only half the job. For a visually impaired user, the invisible layer determines the whole experience. A product page whose alt text has been skipped is a shop window closed to that user.
- Make sure the files you send for translation contain the alt texts as well.
- If you have replaced the images you localised, update the alt text too.
- Tell your team which images are decorative so that no unnecessary description is added.
- After delivery, listen through the content with a screen reader and check it.
If you would like to establish which of these steps you need, you can look at our translation services and get in touch so that we can set the scope together. Accessibility should not be a layer added afterwards but a choice planned from the very start.
Frequently asked questions
What is alt text?
Alt text is the short description written into the alt attribute in an image’s HTML code, which conveys the content of the image in spoken form to a user who cannot see it.
Why is alt text translated?
A screen reader speaks this text aloud in the user’s language. If it is left in the source language, a visually impaired user cannot understand the content, which is why it has to be localised.
What should the alt text of a decorative image be?
For images that are purely decorative the alt attribute is left empty, so that the screen reader skips it and does not wear the user out with an unnecessary description.
Who writes alt text?
Ideally the translator who sees the context with the image in place; if the source alt text is missing, localisation is an opportunity to supply it.
References
- Jiménez-Crespo, M. A. (2024). Localization in Translation. Routledge (Chs. 6/7).
- W3C, Web Content Accessibility Guidelines (WCAG), Success Criterion 1.1.1 Non-text Content.
- Torres del Rey, J. & Morado Vázquez, L. (2019). Accessibility in localisation.