Skip to content

Persian (Farsi) Image to Text

Get editable Persian text from a photo, scan or screenshot. The tool opens with Persian already selected, reads the image in your browser and gives the text back right to left, ready to copy into Word, Google Docs or Telegram.

Add an image with Persian text

Supports JPG, JPEG, PNG, GIF, BMP, WebP, TIFF, HEIC, HEIF, HEICS

Auto detects the writing system and reads it. For better results, pick the language (up to 3 if the text mixes languages).

Drop a photo, scan or screenshot in Persian here

or

Max file size: 10MB

More OCR Tools

Convert different image formats to text with our specialized OCR tools

By task

By language

By file type

What to expect from Persian OCR

Persian is written right to left in the Arabic script, plus four letters of its own: پ چ ژ گ. Letters join inside a word and change shape with their position, and a few differ only by their dots, so image sharpness matters a lot.

Persian specific points:

  • Pick Persian, not Arabic. The Arabic model is not trained on the four Persian letters and may output Arabic forms of ی and ک. With Persian selected, the output uses Persian letters, which matters for search and spell check.
  • Half spaces. Persian uses the zero width non joiner (nim fasele) inside words such as می‌خواهم. In an image it looks like a small gap, so the OCR may output a full space or no space at all. Check prefixes like می and plural endings like ها.
  • Print style. Standard Naskh style type from books, websites and forms reads best. Nastaliq headings and poetry in decorative fonts are much harder and often need correcting.
  • Digits. Persian digits (۱۲۳) and Western digits both appear; check numbers in dates, prices and ID fields.

The result box is set to show right to left text, and the .docx download marks Persian lines as right to left, so the text keeps its order when you copy or open it.

How to extract Persian text from an image

Three steps, no account. Persian is already selected in the language box.

Last updated:

  1. 1

    1. Add the image

    Drop a JPG, JPEG, PNG, GIF, BMP, WebP, TIFF, HEIC, HEIF, HEICS file up to 10 MB, paste a screenshot with Ctrl+V or load it from a link. Crop away anything that is not text.

  2. 2

    2. Extract the text

    Click Extract text. The image is straightened and cleaned up, then read with the Persian model on your device. The first run downloads the model once.

  3. 3

    3. Check and copy

    Fix misread letters and half spaces in the box, then copy the text or download it as .txt or .docx.

Extract Persian text

Tips for better Persian results

Before you extract

  • Keep Persian selected rather than Auto: Auto recognises the script but can settle on Arabic or Urdu
  • For old or faded print, tick Black and white under the language box and extract again
  • Get close enough that dots and the small strokes of گ and ک are clear
  • Use the original photo, not a copy forwarded through a messenger, which is often recompressed

Mixed text

  • Persian with English words? Click Add a language and pick English, so both are read in one run
  • You can combine up to 3 languages, for example Persian and Arabic for a text with Quran quotes
  • Search the result for spaces around می and ها and turn them into half spaces where needed
  • Several pages? Multi Image OCR reads up to 20 images in one batch

What people use it for

Frequently Asked Questions (FAQs)

Our image to text converter is designed to streamline the process of extracting text from images. However, we understand you might have a few questions. To make the most of this tool, we've compiled some frequently asked questions along with helpful explanations to guide you.

Is Farsi the same as Persian here?

Yes. Farsi is the Persian name of the language, and Persian is the option in the language box. It covers text written in Iran and Afghanistan (Dari) in the Persian alphabet. Tajik, written in Cyrillic, is not covered.

Why are some words split or joined wrongly?

Usually because of the half space (nim fasele). In an image it is only a narrow gap, so the OCR sometimes writes a full space or nothing. Fix these in the result box before copying, or in your word processor's search and replace.

Can it read Nastaliq poetry or handwriting?

Not reliably. The Persian model reads standard printed type best. Nastaliq headings, calligraphy and handwriting usually give partial results that need heavy correction. See handwritten to text for handwriting tips.

Is my image uploaded?

No. The Persian model is downloaded to your browser and the image is read on your device, so documents and private chats stay with you.

Latest from Our Blog

Tips, guides, and insights on OCR technology

View All Posts

· Guides

Is Image to Text AI? How OCR Reads Your Photos

Is photo to text AI? Yes, a small neural network reads each line. How Tesseract's LSTM works, how it differs from AI chatbots, and what that means for you.

This tool in other languages