Reuse the useful details
Extract event information or instructions, then verify dates and names.
Turn Ancient Greek writing in photos, screenshots and scans into editable text.
The document language is ready. Change it only when you need to.
Use Future for flexible recognition or Classic for a traditional scan.
Up to 25 MB
PNG, JPG, TIFF, BMP, GIF, WebP or HEIC. The first image or frame is used.
Drag on the page to select an area. Move or resize the box; arrow keys move it and Shift + arrows resize it.
Select the whole page or draw, move and resize a box around the text you need. Each click processes this selection only.
Your result is ready. Download a copy, or return to adjust the settings.
Translate the recognized text, then review and use the result.
Ancient Greek Image OCR is a free online tool that extracts Ancient Greek text from photos, screenshots and scanned images. Upload a JPG, PNG, WebP, TIFF or another supported image, then recognize the whole image or a selected area. Process one image at a time, review and correct the extracted text, and copy it or download it as TXT or Word. Smart Edit and translation tools help you prepare the result for your next task.
Keep the selected language when it matches the document, or choose another recognition language.
Choose Classic or Future, upload an image if needed, and select the whole image or an area.
Click Start, complete verification, check the text and download the format you need.
Image OCR reads the words in a photo, screenshot or scan. Recognize one still image or selected area, compare the text with its source, then correct, format or translate it for your next task.
Extract event information or instructions, then verify dates and names.
Select a passage without retyping the surrounding interface.
Review the wording before editing, translating or saving a new document.
Open one supported image up to 25 MB. Use the whole image or draw a rectangle around the text you need.
Image formats include PNG, JPG, JPEG, TIF, TIFF, BMP, GIF, WebP and HEIC. The image uploads to prepare its preview. Animated and multipage image files use the first frame or page in this OCR workflow.
Move and resize the recognition rectangle to include complete lines. The image editor can help with brightness, contrast and cleanup before you click Start and complete verification. Each request recognizes one image or selected area.
For a later frame in a multipage image, first use a native image tool to select and save that frame, then open the still image in Image OCR. The OCR upload limit still applies to the exported image.
Search 128 recognition languages. Keep the current selection when it matches the source, or choose another document language.
The recognition-language picker describes the words in your source. It is independent of the website interface language. Selecting another recognition language opens its dedicated OCR page.
Choose the language that best matches the passage you want to read. A language selection is not a promise that every word, mixed-language line or historical typeface will be recognized correctly.
Classic includes single-column and multiple-column settings. Future offers another recognition approach for supported languages.
For Classic, choose a column setting that matches the selected content. A focused single-column region can be easier to review than several unrelated text blocks.
Both modes provide editable text and Word output. Classic can also produce a Searchable PDF retaining the selected image area with its recognized text layer. Future's Simple PDF creates a new text layout rather than retaining the original scan design.
Review the source beside the recognized text regardless of the output format. Neither mode guarantees a perfect transcript or an editable reconstruction of the original tables, pictures and typography.
Use the source preview and editable result together. Check important words, numbers and reading order before copying or downloading.
Zoom the source to inspect names, dates, amounts, identifiers and punctuation. Correct the result directly when a letter is wrong or a line break interrupts a sentence.
TXT and Word downloads use the current text. Future's text-only PDF can use the corrected text as well. Later edits do not rewrite the original OCR layer in a Classic Searchable PDF; keep that distinction in mind when choosing a format.
Use the Simple or Advanced editor, local cleanup, history and optional AI actions. Apply Changes returns the edited plain text to the main result.
Advanced editor provides headings, emphasis, lists, alignment, text direction, font choices, colors and highlighting. Simple editor is useful for plain-text corrections. Editing the plain text can remove rich formatting, so choose the workflow that matches your intended output.
Summarize text, Paraphrase text, Fix grammar and Suggest a title are optional server-processed actions. Select a passage for a focused request or leave nothing selected to use the whole editor. A request is limited to 600 words and 12,000 characters; shorten the passage when needed.
Remove accents and vowel marks, normalize number shapes, and use Undo, Redo or Restore to compare changes. These tools are useful for cleanup, but changing spelling or wording is different from transcribing the source. Review any change that could alter meaning.
The main result offers TXT, Word and PDF. Inside Smart edit, direct Word, PDF and HTML downloads keep the editor's supported formatting.
Use TXT for plain text, or the main Word download for an editable document in a new text layout. Choose PDF with the mode distinction in mind: Classic retains the recognized image area, while Future produces text in a new layout.
For headings, lists and other Smart edit formatting, download Word, PDF or HTML directly from that editor. Apply Changes sends plain text back to the main OCR result; it does not turn the original scanned page into editable layout objects.
Download anything you need before reloading the page or starting another file. A working editor is not a saved project, and exports should be opened and checked in the application where you will use them.
Translate text uses the current recognized text, up to 25,000 characters, and lets you review the translated result before applying it.
Correct recognition mistakes first, then open Translate text and choose the target language. The request translates the entire current text in the main result; highlighting a passage there does not limit translation to that selection.
Copy translation copies the translated version without replacing the main result. Use translated text replaces the editable OCR text with the translation. Keep a copy of the original transcript if you need both versions.
Translation can change nuance or misread an ambiguous phrase. Review names, quantities and important instructions, and consult the source when the result is uncertain.
Open one supported image up to 25 MB. Image dimensions, decoding and processing limits also apply. One still image or selected region is recognized per request.
PDF previews and ordinary text editing run in your browser. Images upload after selection to prepare a safe preview. OCR starts after verification. Requested OCR, translation and AI editing use server processing and may use processing providers. Read the Privacy Policy for file handling and storage details.
Inspect accents, breathings and other small marks alongside the base letters.
Keep the Greek passage separate from commentary or parallel translations when selecting text.
Extract Ancient Greek from textbook snapshots and classroom handouts.
Convert scans of lexicon entries or scholia into editable text.
Digitize Ancient Greek quotations in slides, PDFs rendered as images, or screenshots.
A few answers before you get started.
Prepare your imageCompare accents, breathings and iota subscripts with the source, especially in small print. Check line order around notes and commentary, and correct uncertain characters before using the passage as a quotation or study text.
Keep the Greek passage and Latin commentary separate during review. Inspect uncertain letters in the intended script rather than accepting a similar-looking character from another alphabet. Check names, abbreviations and reference labels individually before using the passage in a quotation.
Image OCR accepts PNG, JPG, JPEG, TIF, TIFF, BMP, GIF, WebP and HEIC, one file up to 25 MB. Pixel dimensions and decoding limits also apply. Animated or multipage files use their first frame or page.
Yes. Draw or adjust a recognition rectangle around the passage, including complete lines and punctuation. Click Start and complete verification. Each selected area is a separate recognition request.
No. The current recognition language is already selected. Choose another language when it better matches the text in your document. The searchable list contains 128 recognition languages, separate from the website interface language.
Classic includes single-column and multiple-column settings and can create a Searchable PDF with the selected image and its OCR text layer. Future is available for supported languages and creates a text-only Simple PDF in a new layout. Both provide editable text and Word output.
The main Word download is an editable text document, not a full reconstruction of the source page. Columns, tables, pictures and typefaces may need to be arranged again. Use Smart edit for supported text formatting before its direct Word, PDF or HTML export.
Corrections update the current TXT and Word text, and Future's text-only PDF can use the edited wording. They do not rewrite the original text layer in a Classic Searchable PDF. Use a text export when you need the corrected wording in a new document.
Correct plain text, format it in the Advanced editor, clean up characters, use Undo and Redo, or request a summary, paraphrase, grammar correction or title. Download formatted Word, PDF or HTML directly from Smart edit. Apply Changes returns plain text to the main result.
Yes. Translate text processes the entire current result, up to 25,000 characters. Review it, then choose Copy translation or Use translated text. The latter replaces the main editable text, so keep the original transcript if you need both versions.
Yes. A plausible-looking word or number may still be wrong. Compare names, dates, amounts, identifiers and important wording with the source. Handwriting, unusual fonts, low resolution and complex layouts can need more careful review.
PDF previews and ordinary text editing run in your browser. Image inputs upload for preview preparation. OCR begins after verification; OCR, translation, AI editing and some document exports use server processing and may use processing providers. Read the Privacy Policy for file handling details.
Open your uploaded image in a preparation tool. Return to Image OCR from its result when you are ready.
QUICK VIDEO TUTORIAL
How to extract text from images and save a checked copy
From a picture of words to text you can use.
English narration
Watch the video