DeepL image translation: what it does and how it compares

Yes, DeepL translates images, in two different ways. Its apps read the text in an image and translate it: as text in the desktop apps, laid over the photo in the mobile apps. Its file translation accepts PNG and JPEG files and, according to DeepL's support center, returns a translated image file. TranslateMyImage does one narrower thing: it regenerates the image with the translated text in the original layout. Which one you want depends on what you will do with the result.

What DeepL does with images

In the apps: translated text

DeepL's Translate images page describes the feature in its apps:

• Windows: an Images tab where you drag and drop or upload an image, plus a screen capture tool.
• Mac (macOS 14 or later): drag the image into the text area, or use the screen capture tool.
• iOS and Android: choose the languages, then take a photo or pick one from your gallery, using the buttons below the source text field (DeepL's pages name a Camera and an Image button).

On Windows and Mac, DeepL's support center says the text from the Windows Images tab or from the screen capture tool is extracted, inserted into the source text field and translated as text (Windows images, screen capture); on Windows you can also apply a glossary or the formal/informal feature there. In the mobile apps, the translated text is laid over the original text, with icons to turn the overlay off and copy the translated text (mobile apps). Supported files are BMP, PNG, JPG/JPEG, TIF/TIFF and GIF. DeepL warns that handwritten text is hard to translate accurately, and the free plan is limited to 1,500 characters per day for this feature.

Through file translation: a translated image

DeepL's support center has an article on translating images via file translation. It describes uploading a single PNG or JPEG file: DeepL extracts the text, translates it into any DeepL language and embeds it back into the image, and you download a translated image. The article lists the web translator, the Windows and Mac apps and the API, says the feature is in beta and requires a DeepL account, recommends a minimum resolution of 300 DPI, and states that images inside other documents, such as PDF, Word or PowerPoint files, are not translated. DeepL's file format limits set 3 MB for PNG and JPEG files and allow one image per month without an active DeepL subscription. DeepL's API reference lists JPEG and PNG as "Image (currently in beta)".

What TranslateMyImage does differently

TranslateMyImage uses an AI image model to produce a new image in which the translated text replaces the original, in the same position, while aiming to keep the layout, colors and illustrations. Image translation is the core of the product, so the workflow around it is built for images:

• Modes that change how the text is handled: Marketing (persuasive tone, calls to action), Manga/Comics (speech bubbles, reading flow, sound effects) and Technical Documentation (annotations, precise terms, measurements), plus Default.
• Batch translation on paid plans: up to 10 images per run, each in several target languages, up to 50 outputs, downloaded as a ZIP.
• An API, described in the API documentation and included from the Pro plan up.
• A custom instruction on paid plans, for terminology such as "keep product names in English".

The limits are just as real. Because the image is regenerated, digits, small print and logos can change, so every result needs a check against the original. The uploader translates into 9 languages (English, French, Spanish, German, Italian, Portuguese, Japanese, Korean, Chinese), and the language pages cover 45 target languages. There is no text export, no editing of the translation in the browser, no built-in camera mode, and no Word, PowerPoint or PDF input.

Feature comparison

Choose DeepL when

• You want the words, not a picture. Reading a letter, quoting a screenshot, pasting a translated paragraph into an email: text output is what you need, and the apps give it to you directly.
• Your files are documents. Word, PowerPoint and PDF files go through DeepL's document translation. TranslateMyImage does not accept them.
• You already work in DeepL and want image text to follow the same workflow as the rest of your translations.
• Your target language is not covered by our tool (9 languages in the uploader, 45 through the language pages).

Choose TranslateMyImage when

• The image is the deliverable. An ad, a product image, a social post, a manga page, a diagram for a manual: you need a finished visual in another language, not a list of translated strings.
• The content type matters. Modes for marketing copy, comics and technical documentation change how text is handled.
• You have many images or many languages. Batch runs and the API are designed for that.

If you are unsure, test both on two or three of your real images. For other options, including Google Translate's Images tab and ChatGPT, see our image translators compared, the Google Translate comparison and Can ChatGPT translate images?

Using both on the same project

The two tools do not compete on every job. A common case is a slide deck or a manual: DeepL's document translation handles the text, but screenshots, charts and diagrams embedded as images stay in the original language. Export those images, translate them with the image translator (or with DeepL's own file translation, as its support center suggests), and put them back. The PowerPoint guide walks through that workflow.

DeepL is also a useful second opinion when you review a generated image. Run the original image through DeepL's app, read its text translation, and compare the key terms with the wording in the translated image. Where they differ, ask a native speaker.