Image to Text Converter (OCR)

How this OCR tool works

The tool reads each image directly in your browser using an open-source text-recognition engine (Tesseract). Nothing is sent to a server — only the small language files needed for recognition are downloaded once from a public CDN the first time you use a language.

Plain text vs. layout (approximate)

Plain text gives you the words in reading order with normal line breaks — best for pasting into a document or search. Layout (approximate) tries to keep columns and simple tables lined up by adding extra spacing based on each word's position in the image. It's a best-effort approximation, not a real table: a downloaded .txt file can't hold cell borders or merged cells, so complex tables are still worth double-checking against the original image.

Low-confidence words

Words the engine is unsure about are underlined in the result. Hover over one to see its confidence score. This is usually caused by blur, low resolution, unusual fonts, or handwriting — worth a manual check before you trust the text.

Tips for better accuracy

• Use the sharpest, highest-resolution version of the image you have.
• Crop out anything that isn't text before uploading.
• Use the rotate button if the text isn't upright — recognition is much less accurate on sideways or upside-down text.
• Set the correct language before you start — recognition is far less accurate when it's run in the wrong language.

Limits

To keep things fast and reliable in the browser, each image is capped at MB, batches at images, and MB combined per batch.