What the DeepL image translator can do, and when to use something else

Where can you translate an image with DeepL?

In the apps. DeepL's image page gives the steps for its desktop and mobile apps, not for the website[1]:

Windows app

Open the Images tab and drag in or upload an image, or use screen capture[1]

Mac app

Drag an image into the text field, or use screen capture next to History. macOS 14 or later[1]

iPhone and Android

Tap the Camera button under the source text, then take a photo or pick one from your gallery[1]

Screen capture reads text on anything on your screen, such as a website or a PDF you cannot copy from. It reads 14 source languages on Windows and 9 on Mac[1]. For text in photos, DeepL refers to a separate list of languages for its mobile apps[1].

What does the free plan allow?

Up to 1,500 characters of image translation a day[1]. A street sign or a menu fits easily. A full page of a contract or a birth certificate can run past that.

There is a second limit for personal documents. DeepL's privacy policy says its terms do not allow texts containing personal data of any kind on the free version, only on DeepL Pro[2]. On the free version, DeepL also uses uploaded content for a limited time to train its models[2]. With Pro, DeepL says no one else can access your data and texts are deleted after translation[1].

Which image formats does DeepL accept?

BMP, PNG, JPG, JPEG, TIF, TIFF and GIF[1]. HEIC is not on that list, so save an iPhone photo in HEIC format as JPG before you upload it from a computer.

DeepL's file translation is a separate route. Its document page says it uses OCR to translate scanned PDFs and PNG and JPG images while keeping their original layout[3]. The same page says a free account covers one Word, PowerPoint or PDF file a month, up to 5 MB, and that the other file types need a Pro plan[3].

Do you get a translated image or editable text?

In the apps, DeepL shows the translation of the text it found in the image[1]. DeepL suggests passing it on to DeepL Write to improve the grammar and style[1]. For scanned documents, its PDF page recommends the desktop and mobile apps and says you receive the translation of the text extracted from the file[4].

That suits a menu or a screenshot. For a form, the table structure matters as much as the words. LoreAI reads the photo with OCR and returns a Word file with headings, lists and tables in the original order, so a field label stays next to its value.

How does DeepL handle a multi-page scan?

The image page describes the steps for one image or one screen capture at a time[1]. For a scanned PDF with several pages, DeepL points to its apps and returns the extracted text[4]. On the free plan the 1,500 character daily cap counts across all of it[1].

LoreAI takes the whole scanned PDF in one upload, up to 3 pages on the free document, and returns one Word file.

DeepL image translation and LoreAI side by side

We did not run the same photo through DeepL and LoreAI for this page. Everything about DeepL comes from its own pages on the date in the caption.

When to use LoreAI instead

• The image is a document, such as a certificate, a transcript or an ID, and you need it back as a file you can edit.
• The text runs past 1,500 characters and you do not have DeepL Pro[1].
• The photo contains personal data and you only have DeepL's free version[2].
• You work on a phone and do not want to install an app.

For a sign, a menu or a screenshot, DeepL's apps are quicker. For document files on DeepL, such as typed PDFs and Word files, see the DeepL document translation review. To compare more apps for photos, see the best image translators.

Neither tool gives you a certified translation. LoreAI is not a certified translation service, and the guide to getting a certified translation explains the next step for an authority.

Questions people ask