As of now, it supports OCR processing of the Computer Vision API v1.0.
Feel free to add support for other APIs via pull requests.
Invoke either of the two overloads of the RecognizeAsync method of the JanLenoch.CognitiveServices.Ocr.OcrClient class. The first overload is suited for submitting URIs of publicly accessible images, the second one accepts a byte[] array of an image.
The methods return an OcrResponse object with regions, lines and words. All hierarchically structured, with bounding boxes and coordinates.