tinymodelLocal by nature

Image to Text

From screenshot to something editable. Recognize multilingual text in images.

Files stay on your device No accountEditable text · TXT
PRIVATE WORKSPACEOn your device · PP-OCRv6

A little possibility starts here.

Choose an image to get started. Your preview will appear here.

Private by design
Files and results stay in this workspace. Only application code and model assets are downloaded.

Three small steps.

  1. Choose a clear image and rotate it upright if needed.
  2. Run the multilingual detection and recognition models locally.
  3. Review the text, make corrections, and copy or download it.

A few things to know.

PP-OCRv6 Small

Best with upright, legible text. Complex columns, handwriting, curved text, and tables may need correction. Table structure is not preserved.

The model downloads on first use. Processing speed depends on your browser and device. Review results before relying on them.

Get a little more out of it.