Hosted Microsoft OCR library: Free OCR API

Do you want to test the new Microsoft OCR library Microsoft.Windows.Ocr in an ASP.NET project, a C#/VB.NET desktop (non-Windows Store) app, Chrome extension, or an Android/iOS/Mac/Linux application? Then read on.

The OCR API web service (“OCR-as-a-Service”) is now officially live! A big “thank you” to the many beta testers in the last weeks.

What problem does the OCR API web service solve?

  1. Easy-to use free OCR API
  2. No registration required - perfect for quick prototyping of an OCR app.
  3. Works with images and PDF documents
  4. Each free API key includes [15000]( 25,000 requests/month, more than enough for most.
  5. The OCR quality matches the best commercial solutions.
  6. Hosted in Germany, so the strict European privacy laws apply.

Some background info: Microsoft research released a very good and free OCR library for Windows Runtime as a NuGet package for Windows 8 in 2014. And now, in Windows 10, the OCR library is even part of the operating system. In Win 10 and (soon) Windows Server 2016 the Windows.Media. Ocr namespace for Universal (UWP) apps provides the “OcrEngine” class for optical character recognition (OCR), which replaces the need to install a separate Microsoft OCR library.

The Internet: Not only good for cat pictures, but also for a free OCR Web Service

Microsoft’s OCR engine is easier to use and gives better results than Tesseract.It empowers developers to easily add text recognition capabilities to Windows Phone and Windows Store apps. The catch is exactly that:_ The library works only Windows Phone and Windows Store apps, which is a strong limitation. It can not be used in a classic Win32 desktop application, much less in a web application or iPhone/Android app. Here comes our Cloud OCR API to the rescue:

OCR command line with cURL: One line of code is enough to get started!


The Cloud OCR API provides a REST API as a simple way to parse images and get the text results in JSON. There is no need to fiddle around with a way to reference the Microsoft.Windows.Ocr (WindowsPreview.Media.Ocr.dll ) assembly or the like just for a proof-of-concept. Instead, just consume the OCR web api from any internet-connected device (web app, server, desktop, mobile, iPhone, Android, Mac, Linux, refrigerator…).

Why was this service created?

The OCR API, just like the Online OCR web app and the open-source Copyfish Chrome extension, is a by-product of my optical character recognition experiments. Once the Online OCR web service was live, I received requests for an API. And I needed one anyway for the upcoming Copyfish Chrome extension. So here it is. As with any software, the more people use it, the more robust the solution gets. So enjoy - and please email me with questions and suggestions, I am still actively improving the service.

The use of the OCR API is completely free. Just note that the API is provided “as is” and is not intended for production use. This is not a commercial service.

Update 2016/1/2: The free OCR API service was upgraded. It supports now 100,000 requests per month.