Turn printed words in a picture into editable text
Use image to text when you can see words in a photograph or screenshot but cannot select them. Text recognition can provide a starting point for copying a note, reusing a passage or transcribing a short document.
The image to text converter reads the picture and returns plain text. You can copy the result or download it as a text file, then edit the wording in your own application.
Recognition is not a guarantee of an exact transcription. Review the output beside the source, especially when names, numbers, punctuation or unusual terminology matter.
How to convert image to text
- Choose a JPG, PNG or WebP image and select the language of its printed text.
- Allow the recognition files to load, then run text extraction.
- Review the recognized words and copy or download the corrected text.
The tool works without an account. Recognition takes place in your browser. It may need to download processing and language files on first use, so keep the page open while loading and recognition finish.
Each image must be within 40 MB and 24 megapixels. A batch can contain up to 30 files and 160 MB of combined input. Start with one representative image when you are checking a new language or document type.
Choose the language that matches the page
Language selection helps the recognizer interpret letter shapes and word patterns. Choose English for English text, or the matching option for another supported language.
Available choices include Simplified Chinese, Chinese with English, Traditional Chinese with English, Japanese with English, Korean with English, French, German, Spanish and Portuguese. For a mixed-language source, use an available combined option that fits the content.
If the text is misread, check the selected language before repeatedly processing the same image. A clear source with the wrong language setting can still produce an unreliable result.
Give the text enough space and contrast
Use a sharp picture where the letters are large enough to read. Tiny text in a full-screen screenshot may need a tighter crop around the relevant paragraph.
Keep complete letters and lines in that crop. Cutting off the edge of a character can turn one word into another or remove punctuation that changes the meaning.
Rotate a sideways image before recognition. Aim for an evenly lit source with minimal glare, shadows and perspective distortion. If you can retake a blurry document photograph, that may help more than changing settings on an unreadable picture.
Expect plain text rather than a recreated document
Image to text returns wording, not a Word document that duplicates the original page. It does not promise to preserve fonts, columns, table cells, images or precise page layout.
A table may need manual reconstruction after the words are recognized. A two-column article may need its reading order corrected. Headings and paragraphs can require extra line breaks or spacing.
If you need to preserve the appearance of several photographed pages, save the pictures as PDF. If you need to edit or reuse the wording, recognized plain text is the more relevant output.
Review the details that are easy to confuse
Check numbers separately from surrounding prose. Similar shapes can be mistaken for one another, such as a letter and a digit. A misplaced decimal point or missing minus sign can be more consequential than an obvious spelling error.
Compare names, dates, item codes, addresses and amounts directly with the image. Do not rely only on whether a sentence reads fluently: an incorrect word can still look plausible in context.
For a long source, review in small sections. Keep the image and extracted text side by side so you can track what has and has not been checked.
Know when another source is better
If text is already selectable in a document or webpage, copying it directly avoids recognition errors. Image to text is most useful when the words are trapped in pixels.
Handwriting, decorative lettering, curved text and very low-quality scans can be difficult to recognize. Treat the output as a draft and be prepared to type uncertain sections yourself.
This tool reads text; it does not translate it. A source in another language remains in that language unless you take the extracted text to a separate translation tool.
Convert image to text: common questions
Can I copy the recognized text?
Yes. Review the result, then copy it or download a plain-text file for editing in another application.
Does this produce a Word document?
No. The output is plain text. It does not recreate a DOCX file or preserve the full visual layout of the source.
Can I recognize text from a screenshot?
Yes. Clear screenshots with readable printed text are suitable inputs. Crop away unrelated regions if they distract from the passage you need.
Can I use a scanned PDF?
This page accepts images. Render the relevant PDF pages with PDF to JPG, then recognize their text here. If the PDF already contains selectable text, copying it directly is preferable.
Will handwriting be accurate?
Do not count on it. This tool works best with printed text, and handwriting can produce substantial errors.
Why does the first run take longer?
Your browser may need to load the recognition engine and selected language files. Keep the tab open until loading and text extraction finish.