OCR, In-Browser

Read the text in any image, running entirely on your device.

A two-stage text detection and recognition pipeline, exported to ONNX and executed locally with ONNX Runtime Web — accelerated by WebGPU where available, falling back to WebAssembly otherwise. Nothing you upload is sent anywhere.

Object Detection Instance Segmentation Semantic Segmentation Pose Estimation Depth Estimation OCR Background Removal Visual Search
Images are processed locally in your browser and are not uploaded to a server.
Loading…
Unable to load the model.

Image

Drag & drop an image, or click to browse
JPEG, PNG, or WebP — text on a sign, receipt, or page

Results

Run an image to see recognized text.

Performance
TotalRun an image to see metrics

Controls

Detection confidence0.40