STEP 01Step 1 — Upload your image
Drag a JPG, PNG or WEBP file into the AI image translator (up to 16 MB). Photos, screenshots, manga pages and scanned PDFs all work. The clearer the text, the better the translation.
AI IMAGE TRANSLATOR · FREE · 30+ LANGUAGES
AI image translator — translate image to English with layout-aware AI rendering.
Translate text in photos, manga, menus and PDF pages into 32 languages. The selected image-editing model attempts to match the original layout, typography and background; always preview the generated result.
★ 6 image models·Up to 20 images per batch·32 languages supported·Temporary links · files expire after about 20h
JA → EN
JP MANGA → EN
ZH → ENImage translation workspace
A translation desk, not a black box.
Translation proof
One image, carefully reset.
Add a menu, label, sign or screenshot to compare the original and translated artwork.
Production queue
Ready to estimate
Overall progress0%
Your selected files and their live status will appear here.
Recent translations
Metadata stays for 90 days; image files expire after about 20 hours.
Up to 20 images per batchFailed items refunded32 translation languages6 image-editing modelsLayout-preserving translationFiles expire after about 20hNo Imya watermark on outputsManga, menus, signs and PDFsFolder, ZIP and PDF workflowsUp to 20 images per batchFailed items refunded32 translation languages6 image-editing modelsLayout-preserving translationFiles expire after about 20hNo Imya watermark on outputsManga, menus, signs and PDFsFolder, ZIP and PDF workflows
WHO USES IT
Who reads with the AI image translator
The AI image translator sends each image through a generative image-editing model that tries to replace visible text while keeping the surrounding artwork recognizable. Results vary on handwriting, tiny text and dense layouts, so the before-and-after preview stays part of the workflow.
01 · TRAVELERS & DINERS
Read every menu, sign and street in the wild
Use travel photos for menus, station signs, storefronts and labels. The model attempts to place translated wording back into the same visual regions, while the preview lets you check prices, times and directions before relying on the result.
- Read handwritten chalkboards, painted signs and printed menus
- Keep prices aligned with dishes and times glued to platforms
- Auto-detect the source language across 30+ scripts
- Save the translated photo to share with travel companions
- Use signup and daily check-in credits · no app install
02 · MANGA & WEBTOON READERS
Read every chapter, even the ones nobody translated
Drop a manga page or webtoon into the translator for a generative first pass. It attempts to keep bubbles, panels and artwork in place, but stylized sound effects and small lettering may need a retry or manual polish.
- Detect and re-render speech bubbles in matching fonts
- Preserve sound effects, panel order and gutters
- Handle vertical webtoons and horizontal manga layouts
- Work on doujinshi, indie comics and decades-old scans
- Cut scanlation typesetting time from hours to minutes
03 · ARCHIVISTS & RESEARCHERS
Read the documents OCR engines give up on
Use scanned letters, postcards and journals as a visual translation aid. Cursive, faded ink and damaged paper are difficult inputs for any model, so treat the generated page as a draft and verify important wording against the original.
- Handle cursive, calligraphy and pre-war fonts
- Read faded ink, water damage and folded paper
- Produce side-by-side originals and English transcriptions
- Cover French, German, Polish, Russian, Yiddish, Italian and more
- Files use temporary public links and expire after about 20h
THREE STEPS
How to translate an image with AI
No software to install. Guests can translate one GPT Image 2 image per day; signed-in users can run batch and PDF jobs. Model estimates are usually 25–60 seconds, and queues or complex images may take longer.
STEP 01
STEP 02Step 2 — Pick a target language
Choose from 30+ target languages — English, Chinese, Japanese, Korean, Spanish, French, German, Arabic and more. The AI image translator auto-detects the source language, or set it manually for mixed-script pages.
STEP 03Step 3 — Download the translated image
Preview the before and after side by side, then export as the original format, JPEG, PNG or WebP. Match mode resizes the export to the source pixel dimensions; generated typography and layout can still vary.
LANGUAGES
Translate image to English, Japanese, Chinese, Korean and 30+ more
The AI image translator handles every major script — Latin, Cyrillic, CJK, Arabic, Devanagari, Thai, Hebrew. Hover any pill below to see its English name.
- EnglishEnglish
- 简体中文Chinese (Simplified)
- 繁體中文Chinese (Traditional)
- 日本語Japanese
- 한국어Korean
- EspañolSpanish
- FrançaisFrench
- DeutschGerman
- ItalianoItalian
- PortuguêsPortuguese
- NederlandsDutch
- РусскийRussian
- العربيةArabic
- हिन्दीHindi
- বাংলাBengali
- Tiếng ViệtVietnamese
- ภาษาไทยThai
- Bahasa IndonesiaIndonesian
- Bahasa MelayuMalay
- TürkçeTurkish
- PolskiPolish
- УкраїнськаUkrainian
- עבריתHebrew
- TagalogTagalog
- فارسیPersian
- اردوUrdu
- SvenskaSwedish
- DanskDanish
- SuomiFinnish
- ČeštinaCzech
- RomânăRomanian
- ΕλληνικάGreek
Japanese → English
Chinese → English
Korean → English
Arabic → English
FEATURES
A complete image translation workflow
Single images, folders and PDF pages all use the same visible queue, server pricing and per-item billing rules.
Translate up to 20 images at once
Drop a folder. The batch AI image translator queues every file, shows live progress, and outputs a zip of translated images keeping the same filenames.
Unpacked Chrome extension included
A Manifest V3 extension is included for manual installation. It supports right-click translation and a temporary in-page result overlay, but it is not yet published in the Chrome Web Store.
Compare before and after
Inspect the source and generated image side by side or with a split view. Retry difficult pages when small text, handwriting or dense layouts need another pass.
PRICING
One free guest translation + credit-based runs
Guests get one GPT Image 2 translation per day. Signed-in runs are priced per image from the live server registry, and final failures refund only the affected items.
No Imya watermark · batch credits add per image · failed items refunded.
See pricing →FAQ
AI image translator — frequently asked questions
Is the AI image translator really free?
Guests receive one free GPT Image 2 translation per day with no card. Sign in to use batch, PDF and all enabled models; signed-in jobs charge the live per-image credit price. Translator outputs do not add an Imya watermark.
Can the AI image translator handle manga and comics?
Yes — manga is one of the most common use cases. The AI image translator detects speech bubbles, sound effects and panel order, then renders English dialogue inside the same bubbles. It is widely used by scanlation teams who want a head start on cleaning and typesetting.
What languages does the AI image translator support?
30+ source and target languages including English, Chinese (Simplified and Traditional), Japanese, Korean, Spanish, French, German, Portuguese, Italian, Russian, Arabic, Hindi, Bengali, Vietnamese, Thai, Indonesian, Turkish, Polish, Ukrainian, Hebrew and Tagalog.
Does the AI image translator work on PDFs?
Yes. Signed-in users can load a PDF of up to 20 pages. Pages are prepared locally, translated one by one as images, then rebuilt into a PDF. Diagrams remain part of the page image, but generative results can vary.
Will the AI image translator keep the original layout?
The prompt asks the selected image-editing model to preserve composition, text placement, typography and non-text artwork as closely as possible. This is not a pixel-perfect guarantee, so use the before-and-after preview and retry difficult pages.
Is there an AI image translator Chrome extension?
A Manifest V3 developer preview is included in the project for manual, unpacked installation. It supports right-click translation and an in-page overlay, but it has not yet been published in the Chrome Web Store.
Can I batch translate images?
Yes. The batch AI image translator processes up to 20 images at once and returns a zip with the same filenames. Useful for product catalogs, manga chapters and scanned PDFs.
How accurate is the AI image translation versus Google Translate?
The tools use different workflows, so there is no universal winner. Imya asks a generative image-editing model to rebuild translated text in the image; always verify names, prices and important wording, especially with handwriting or small text.
Is my image private?
PDF pages are prepared locally, but images are uploaded over HTTPS to randomized public R2 object URLs so the KIE model can process them. Source and result files are scheduled for deletion after about 20 hours; task metadata is retained for 90 days. Do not upload material that requires access-controlled storage.
Is there an open-source AI image translator alternative?
Several open-source AI image translator projects exist on GitHub, mostly for manga (manga-image-translator). They are great for hobbyists but require local GPU setup. Imya.ai is the zero-install online AI image translator if you just want to translate images right now.
EN — open access
Translate your first image — free.
No card for the daily guest image. Choose a language, follow live progress and review the generated result before downloading.
Translate my image →