Image to Text (OCR) Guide
Extract Korean text from images locally with WebGPU acceleration.
Extract Korean, English, and number text from photos, scans, and screenshots without sending the image to a server.
How to use
Runtime
A Korean PP-OCRv5 detection and recognition pair runs through PaddleOCR.js and ONNX Runtime Web. WebGPU is selected automatically when the browser supports it; the version-pinned WASM backend is used as a fallback.
The first run downloads about 18 MB of model assets. Keep the page open while the models are cached. Accuracy depends on image focus, lighting, and text layout.
Input limits and expected results
- One PNG, JPG, WebP, or GIF image up to 10 MB.
- Images above the safe pixel budget are reduced before inference; extreme dimensions or prepared images above the internal budget are rejected.
- The recognition model is primarily for Korean, English, and numbers. Handwriting, decorative fonts, vertical text, tables, formulas, and unsupported scripts may be missed or placed in the wrong reading order.
OCR output is a draft, not a verified transcript. Compare names, dates, totals, account numbers, medical text, and legal wording with the source image before relying on it.
Troubleshooting
Crop away unrelated areas, rotate the image upright, increase contrast without clipping characters, and use the original resolution when possible. If the model cannot load, check that content blockers or a restricted network are not blocking the documented model host, then reload once. WebGPU is an acceleration path, not a requirement; the WASM fallback can be slower.
Privacy
OCR runs in a dedicated browser worker. Image pixels are not sent to an OCR API or Aiviko processing server. Model and runtime assets are downloaded separately.
Limits and troubleshooting
The interface can be displayed in five languages, but this OCR model recognizes Korean, English and numbers. It does not provide a dedicated Japanese, Chinese or Spanish recognition model. For PDF or HEIC input, first create a PNG or JPG with the file converter. Preserve the original and compare the extracted text before reusing it.
Frequently asked questions
Does it recognize handwritten text?
This tool uses a model tuned for printed Korean, English, and number text. Cursive handwriting, decorative display fonts, vertical text, and complex formulas can have much lower recognition accuracy.
Can I run OCR directly on a PDF or HEIC file?
Direct PDF or HEIC upload is not supported in the OCR tool. Convert your pages to PNG or JPG first with the PDF to PNG or HEIC to JPG workflows, then run OCR on the exported image.
Are my uploaded images stored on a server?
No. Image decoding and text recognition happen inside a dedicated browser Web Worker on your device, and image pixels are never transmitted to an Aiviko server.
Why does the first run take longer?
On the first run, the browser downloads about 18 MB of PP-OCRv5 model and runtime assets and caches them locally. Subsequent runs in the same browser reuse the cached files and start much faster.
Related tools
Quick answer and specifications
The core workflow runs in this browser. Check the supported formats, published limits, and any on-demand model download before you start.
- Inputs
- JPG, PNG, WebP, & GIF
- Outputs
- text & Markdown
- Processing
- On this device, in the browser
- Model languages
- Korean and English
- Per-file limit
- 10 MB
- Files per run
- 1
- On-demand model download
- 18 MB
- Last verified
The interface language does not determine the model language. Recognition and pronunciation quality vary; check the result.