Capture
Take a photo or select an image from your device.
No. Text recognition runs locally in your browser without uploading the photo. Only the recognized paragraph text is sent to the speech generation service.
Use a sharp, evenly lit photo with upright text and little background clutter. Cropping the image to the page usually improves recognition.
The browser downloads the OCR engine and the selected language model on first use. Later scans can reuse cached files and are usually faster.
Snap to Read supports English, Simplified Chinese, Spanish, French, German, Japanese, Korean, Italian, Portuguese, Russian, and Arabic. Before choosing an image, select the language that matches its text and pick an available voice.
Yes. Sign in before taking a photo or choosing an image. Creating audio uses your account's text-to-speech credits, so make sure you have enough credits for all sections.
Yes. Once an audio section is ready, use the download button beside its player to save it as an MP3. Download the sections you want to keep before starting over, choosing another image, or leaving the page.