Read a picture, in your browser
PP-OCRv6 finds the lines and reads them with tract, on the CPU or on WebGPU. Nothing is uploaded. Tiny, Small and Medium are fetched from Hugging Face the first time and kept in this browser.
Loading the sample…
Sample page from the PP-OCRv6 model card. Detection and recognition weights: PaddlePaddle ONNX, Apache-2.0. The graphs run in tract, compiled to WebAssembly. Source code