Smart tools and reading

Local background removal

Background removal runs a model in your browser. The model downloads on first use; your image is not sent to a server for recognition. Results depend on the subject and background. Hair, transparent objects and complex edges may need further editing.

PDF text recognition

PDF reading mode can extract existing text or use OCR for scanned pages. Select the pages and recognition language before processing. A batch can include up to 20 OCR pages. Clear, upright scans generally work better.

Exporting a reading PDF reflows the text. Check names, numbers, tables and formulas carefully. OCR output is not guaranteed to be an exact copy of the source.

On-device speech

Read aloud uses local voices installed on your device. Languages and voices depend on the system. If a voice is missing, install it in device settings first.

Immersive reading, focus reading and reading fonts offer ways to adjust spacing, fonts and how you read.

Downloads and waiting

The first engine or model download needs a connection; file processing then runs on your device. Retry after checking the network if loading fails. If memory is limited, try a smaller image or fewer PDF pages.