Read text from an image or a PDF
Two different readers
OCR has a local reader and an optional AI reader. Local OCR runs in the browser. It can misread small or unclear type. It can place Latin words on an invisible text layer of a PDF. Chinese, Japanese, and Korean text should be copied from the recognized result rather than assumed to be on that layer. AI mode sends the content you selected to the provider you configured in Settings. The key stays in this browser. NexTool does not receive the image. Read that provider’s policy before you send a private scan. If you only needed to hide a number, use redact instead of OCR. If you only needed a smaller file, use compress.
The file stays in this browser. NexTool does not upload it and does not keep a copy after you leave the page. Download the result and check it before you send it on. If the tab runs out of memory, the job stops and the original is still the one on your computer.