HeyDreaming

Free · processed in your browser · account only to save the .txt

Extract Text from an Image

Loading the tool…

Turn a screenshot, scanned page, or photo of printed English text into editable plain text you can copy or download as a .txt file. Recognition runs in your browser, so your image stays on your device.

How do I extract text from an image in my browser?

Drop a JPG, PNG, WebP, AVIF, HEIC (.heic/.heif), or BMP image up to 20 MB into the box above—or paste a screenshot with Ctrl+V or Cmd+V—and press Read text. The text is read in your browser, so the image is not uploaded (the first run takes a little longer). The tool reads the words, flags any low-confidence words to check, and lets you edit, copy, or download the result as a .txt file.

How it works

01

Add or paste an image

Drop a JPG, PNG, WebP, AVIF, HEIC/HEIF, or BMP file up to 20 MB, or press Ctrl+V / Cmd+V to paste a screenshot from your clipboard.

02

Choose layout options

Pick Auto, Single block, or Single line under Page layout, and optionally check Join lines into paragraphs.

03

Read, check, and export

Press Read text, review any words listed under Check these words, edit in the text box if needed, then press Copy or download the .txt file. Reading and copying need no account; the .txt download needs a free HeyDreaming account.

How in-browser text recognition prepares your image

Optical character recognition (OCR) runs locally in your browser and reads English text; the first run takes a little longer. Before reading, the image is drawn onto a white canvas so dark letters on a transparent PNG have clean contrast.

Very small screenshots and oversized camera photos both hurt recognition accuracy, so the tool scales the canvas in two steps: first, if the short side is under 1000 px, it enlarges the image toward a 1000 px short side (at most 3×); second, if the resulting long side exceeds 3000 px, the whole image is scaled down so the long side is at most 3000 px. That cap always wins—for example, a 720×4000 px strip scales to 540×3000 px.

Cleaning lines and checking uncertain words

Raw OCR output often contains trailing spaces, words split across lines with a hyphen at the margin, and hard line breaks at the end of every printed line. The cleaner automatically trims each line, rejoins hyphenated line-end words, and collapses repeated blank lines. Checking Join lines into paragraphs merges lines within each paragraph with spaces while keeping blank lines between paragraphs; you can toggle it after a run to reformat unedited text on the spot.

After reading, the tool shows an overall confidence rating (high at 85%+, medium at 60–84%, or low below 60%) and lists up to 25 unique words that scored below 60% confidence so you know exactly which names, numbers, or punctuation marks to proofread in the text box.

What this tool does not do

It does not read non-English languages, preserve multi-column page formatting or tables as Word/PDF layouts, or reliably transcribe cursive handwriting.

See all free tools

Keep exploring

Related free tools

Questions

Image to Text: FAQ

Is my screenshot or document photo uploaded?

No. The text is read in your browser, so your image stays on your device and is not uploaded. The page sends usage events to Google Analytics, such as which tool ran, but never your image or extracted text.

When should I change Page layout from Auto to Single block or Single line?

Leave Page layout on Auto for full pages, receipts, and screenshots with headings. Switch to Single block for a cropped paragraph or code snippet, or Single line when the image contains only one horizontal line of text such as a serial number or title.

Why were some characters misread?

Text recognition works best on sharp, upright, printed English text with strong contrast against the background. Handwriting, decorative fonts, skewed camera angles, and low-resolution crops often produce errors—check the flagged low-confidence words and edit them directly in the text area.

Can I paste a screenshot without saving it as a file first?

Yes. Copy any screenshot or image to your clipboard and press Ctrl+V (Windows/Linux) or Cmd+V (Mac) anywhere outside an active text input to load it straight into the reader.

What image formats and file sizes are supported?

You can open one JPG, PNG, WebP, AVIF, HEIC (.heic or .heif), or BMP image up to 20 MB.