Free Image to Text Converter nothing stored

Get copyable words from a screenshot or scanned page

Optical character recognition, or OCR, reads visible letters and returns text you can select. Use it for a captured message, a photographed book passage or a printed receipt. It reads existing words rather than writing a description of the scene.

The result pane shows the recognized text and a Copy text action. Paste the result into your own editor to correct it; the preview itself is read-only. Saving produces a plain .txt file, not a formatted document.

Match the recognition language and reading layout

Language starts at English. The available languages are English, Spanish, French, German, Portuguese, Italian, Hindi, Chinese (Simplified), Japanese or Arabic. Choose the language actually printed in the picture; there is no automatic language detection or translation step. For a mixed-language page, repeat the extraction with the other supported language and compare the results.

Layout offers Automatic, Single block or Scattered text, starting with Automatic. The choices suit a full page, a uniform paragraph or receipt, and words spread around signs or app screens respectively. The scattered layout finds labels without promising a reading order. If columns come back interleaved, try a different layout or crop individual passages before adding them.

Give the reader clear, upright letters

The engine enlarges small sources, stretches contrast and smooths grainy images before reading. Clean screenshots skip the grain smoothing. These steps cannot reliably recover lettering hidden by glare, heavy blur or a fold in the page.

Use a closer, straight-on photograph with even light. Rotate sideways pages before extraction, and crop unrelated surroundings in your image editor. Retake a blurry source when possible. Check names, decimal points and serial numbers against the original rather than trusting a plausible-looking sentence.

Check missing words as well as the confidence score

Uncertain words and weak lines are filtered out before the result appears. That avoids many fragments guessed from background texture, but it can also omit real short labels or faint text. A message that no text was found means nothing survived recognition and filtering.

The confidence percentage averages the words that were kept. It is not a measured accuracy rate for the whole document and does not prove every line was captured. Compare the beginning and end of each passage, then retry with the correct language, another layout or a sharper source.

Handwriting and table cells need a separate check

Printed and on-screen lettering is the intended input. Neat handwritten capitals may work, but cursive notes and decorative lettering are unreliable. Do not assume a readable-looking result is a faithful transcription of handwriting.

A table can yield words and numbers, but the output has no spreadsheet cells, formulas or preserved column widths. Paste into a spreadsheet only after checking which values belong together. A photograph of a page also stays an image input: this converter does not accept a multipage PDF or create a searchable PDF.

Load the reader while keeping the document on your device

The browser fetches the recognition code and the selected language data, then reads the image locally. Those downloads are different from uploading your document. The engine sends the working image to a local browser worker, not a remote recognition service.

The initial run includes loading that reader. Later runs with the same language reuse it while it remains available; changing the language replaces the active recognizer. Phones and computers vary in available memory and processing time, so no fixed completion time is promised. Keep a connection available for initial loading or a language change.

Take words out or replace lettering inside the picture

This converter gives you the text while leaving the source picture alone. It does not erase lettering, change a sign or translate a menu in place. After extraction, correct the transcript in your document or pass it to your preferred translator.

For different words in the original scene, use the linked Photo Text Changer. That separate paid editor changes lettering inside an image. A practical sequence is to read the source here, check the wording, then bring the intended replacement to the editor.

How do you use the image to text?

1. Choose a readable source

   Add a JPG, PNG, WebP or HEIC screenshot, document photo or scan with the lettering upright and in focus.

2. Set the reading controls

   Choose Language to match the source, then set Layout for the arrangement of the words.

3. Extract and proofread

   Run extraction, compare the returned lines with the picture and copy the text into an editor for corrections.

4. Save your result

   2 free downloads, 8 more after sign-in.

When is the free image to text enough?

Image to Text: questions and answers