Upload the whole set in one batch. OCR reads each image, AI translates them all against a shared glossary, and you download the full set together — with terms identical from the first file to the last.
Yes — you can translate many images in a single job instead of one at a time.
Trusted by 1,000+ enterprise brands
To translate multiple images at once, you upload them as a single batch to a translation platform instead of processing each file individually. In Smartcat, every image in the batch goes through the same pipeline:
OCR detects the text regions in each file.
Translation converts that text into your chosen languages.
Rebuild places the translated text back into each image's original layout.
What makes a batch different from repeated single uploads is shared context. All images in the batch use one glossary, so brand names and recurring terms are translated identically in every file, and one translation memory, so a sentence repeated across twenty images is translated once and reused.
One batch can target several languages at the same time, and the finished set downloads together, each image in its original format.
Follow these five steps to translate a whole set of images in one batch:
1
Select one or multiple image files and upload them together — PNG, JPG, JPEG, TIFF, BMP and more. Each image can be up to 30 MB for OCR text extraction, and there is no cap on how many files one project holds: batches of hundreds of files are supported, because the constraint is per-file size, not file count.
2
Pick your target languages. One batch can fan out to every target language at once, so there is no re-uploading per language and no separate job per market.
3
Our image translation agent extracts the embedded text from every file in the batch and translates it against your glossary and translation memory, then rebuilds each image with the translated text in its original layout.
4
Spot-check what matters. Open any image in the set, edit a text box, or route high-visibility files to your own team or to Marketplace reviewers — everyone works inside the same project, seeing each image with its translation in place.
5
Download the set: same filenames, same formats, new languages. Our image translator places the text back into the original layout, so the files are ready to publish without extra design work.
We love having total visibility over our entire translation lifecycle on our Smartcat workspace due to the centralized nature of the platform and being able to gather all the linguistic assets we need, including translation memories and glossaries.
”Explore case study →
1 glossary
The same term in image 1 and image 100
Load your glossary before the batch runs and our translation AI agent enforces it across every file — product names, feature names, legal lines. Consistency stops depending on whoever translated which image.
1 memory
Repeated text translated once, then remembered
Translation memory recognizes the tagline that appears on 40 banners and reuses the approved translation instead of re-translating — and re-billing — it 40 times. The more repetitive your image set, the more this matters.
1 upload
Every language from the same batch
A batch into six languages is still one job. You review by language or by image, instead of re-running the whole process six times.
Consumer translation apps process a single image per pass. That is fine for reading a menu. It falls apart the day you have 60 product images and 4 target languages — that is 240 rounds of upload, screenshot, save, rename. The results land as overlays you can read, not files you can publish.
The slower damage is inconsistency. When each image is translated in its own little session, nothing connects them: the return-policy line phrased one way on image 3 comes back phrased differently on image 58, and the product name your brand team fought for drifts across the catalog. Nobody catches it until a reseller does.
Learn more about AI document translation
for ease of setup
for ease of use
enterprise brands
average rating on G2
Book a 1:1 walkthrough with a specialist and bring your own image set. No commitment.
2-3 days
Translation turnaround, down from 10 days
See how Smith+Nephew shortened a 10-day translation cycle to two to three days.
Up to 70%
Cost savings
Stanley Black & Decker cut translation spending while raising quality.
31 hours
Of work time saved monthly
Babbel’s marketing and L&D teams reclaimed 31 hours a month.
Drop in every file, pick the languages, and download a set that reads consistently from the first image to the last. Free for 15 days with 15,000 Smartwords — full access to translation capabilities, no credit card.
Smartcat is a language AI platform that automates translation and content workflows, combining AI agents, human experts and integrations in one workspace — so a set of files goes out in every language you need from a single job.
A batch image translator lets you upload many image files as one job instead of translating them one at a time. Every image in the batch runs through the same pipeline:
Every image also shares one glossary and one translation memory, so recurring terms come back identical across the whole set.
Our Image Translation Agent is a multiple image translator that detects and translates text inside image files. It then rebuilds the image automatically with the translated content. You don't need design tools. This helps you adapt visuals faster, reduce manual work, and ensure brand consistency.
This is how to translate multiple images at once. Upload any image file — JPG, PNG or other formats — into Smartcat. Choose your target languages. The Image Translation Agent will detect and translate all text. You can review and edit on the platform, then export the fully localized image.
No. Google Lens and the Google Translate app process one image per pass, and they give you an on-screen overlay rather than a translated file.
For reading a single image right now, that is exactly right and faster than any platform. For producing a localized image set you can publish — 60 images into 4 languages, with the same term used identically across the set — you need batch processing and files back.
Because remaking every visual per market is the work being replaced: OCR reads the text, AI translates it against your glossary, and each file is rebuilt in its original layout — so a 60-image set into four languages is one job instead of 240 manual passes.
Your edits feed back into the translation memory and glossary, so the terminology approved on one batch is reused on the next.
Add them to a glossary before the batch runs. Every file in the batch is translated against it, so approved terms are used everywhere, and QA flags any segment where they were not.
This is also where a batch can fall short: consistency is only as good as the glossary you load first, so a batch run with an empty glossary can still drift between related images. Spend ten minutes on terms before you spend one minute uploading. Learn more about image translation
Smartcat accepts JPG, JPEG, PNG, TIFF, BMP, GIF and a number of less common image formats for OCR text extraction, alongside the document formats. You can check the current list of supported formats in the help center. Note that an animated GIF is translated from its first frame only.
Yes. Smartcat supports over 280 languages, including right-to-left scripts and regional dialects. You can localize campaign visuals, infographics, and training content for every market.
Yes — layout, text styling and image quality are preserved, and each file exports in its original format.
Specialized fonts are substituted with close alternatives where the original does not cover the target script, so give complex or heavily styled designs a visual check before you publish.
There is no hard cap on the number of files. Smartcat handles projects of any size, from a single file to batch uploads of hundreds of files, and the constraint is per-file size rather than file count:
If a very large batch struggles, the documented remedy is to upload it in smaller batches.
Do budget human time for review, though: spot-checking hundreds of translated images is still work. Plan real review for high-visibility assets and spot-checks for the rest.
Not for loose images. ZIP is supported as a package format — IDML, DITA, Articulate Rise SCORM, Articulate Storyline and similar packaged content, where Smartcat detects the package type automatically — but not as a generic folder upload. Select the image files themselves and upload them together.
No. When you set up an image job you can either let Smartcat detect the source language or set it yourself per file, so a batch that mixes source languages is workable. Setting it yourself is the safer route when a language has near-identical written variants you care about distinguishing.
Yes. You choose all target languages when you set up the job, and the batch fans out to each of them — review and download per language, without re-uploading anything.
Smartcat is SOC 2 Type II compliant, with encryption in transit and at rest, strict access controls and isolated workspaces. Your content is not shared or used to train third-party models.
Low-resolution or heavily stylized files can fail text detection, and output quality tracks input quality — the better the input, the better the result. If a file's format is not supported, an error appears at upload and the file is not added to the batch.
Image translation is metered per image, not per word: translating the text inside an image via OCR costs a flat 1,000 Smartwords per image, however much or little text that image holds. So a batch's cost tracks its file count — worth knowing before you upload 400 near-empty thumbnails. Ordinary document translation works the other way round, at 1 Smartword per source word. Details are on the pricing page.
Yes — three things to check before you build against it:
The API accepts the same OCR image formats as the platform — JPG/JPEG, TIFF, BMP, PNG, GIF, DJVU, DCX, PCX, JP2, JPC, JFIF, JB2 and AI — so a batch you could upload in the browser is a batch you can submit programmatically. Book a demo for a walkthrough.
Middlebury Institute. (2024). Eight key insights on AI and the future of translation and interpretation.
Mayer, Hannah, et al. (2025). “Superagency in the Workplace: Empowering People to Unlock AI’s Full Potential.” McKinsey & Company.
World Economic Forum. (2025, March). How AI agents are unlocking new business value in 2025.