Sikuli X 1.0 RC1 Released!

Written for Sikuli by Tsung-Hsiang Chang on 2010-12-23

Sikuli X is a new experimental branch of Sikuli. (X stands for eXperimental.)
Many new features and bug fixes are included in this release. All old users of Sikuli 0.9 or 0.10 are recommended to upgrade to X. However, please keep in mind some features are still in experimental, e.g. text recognition and the new API to get the bound of any windows, which means they may not work well or not support all platforms yet.

Download Sikuli X at To learn how to use the new features of Sikuli X, please read the new "The Complete Guide to Sikuli X"


 - New computer vision engine - faster and more reliable.
 - Text recognition and matching (EXPERIMENTAL)
      - find("OK") returns all regions with a "OK" label.
      - region.text() returns the text in the region.
 - Screenshot naming - screenshots can be automatically named with timestamps, the text in them, or manually input.
 - Supports remote images
        e.g. click("")
 - Image search path (SIKULI_IMAGE_PATH) - images can be stored anywhere you like.
 - .sikuli source can be imported as a module.
 - New App class replaces the old openApp, switchApp, closeApp functions
      -, App.close(), App.focus()
      - App.window() returns the bound of the app window as a Region, so you can restrict following actions within that region. (EXPERIMENTAL: Windows and Mac only)
 - Beautified "Run in slow motion" mode.
 - Smooth mouse movement.
 - Better capture mode on Mac (supports multi-screens, no flicker anymore)
 - More special keys are supported (PrintScreen, Num Pad, CapsLock...)
 - New region highlighting API: region.highlight(). (EXPERIMENTAL: Windows and Mac only)
 - New Mouse API: wheel(target, WHEEL_UP/WHEEL_DOWN, steps) for scrolling the mouse wheel.
 - New multi-lingual interface translation: Bulgarian, French, Dutch, Polish, Japanese, Simplified Chinese.


 - command line arguments works with the --test mode.
 - multi-screen actions when the primary screen is not at (0,0)
 - (Linux) open an application in background by default
 - (Windows) fixed a bug in setting the screen shot hot key
 - and much more..


 - Text recognition and matching is implemented with the Tesseract OCR engine, which was
   originally designed for recognizing scanned documents. Please note the OCR technology
   is not perfect, especially for screen text. We believe there is still much space for tuning its
   performance and we are working hard to improve it. Please look forward to it.

 - The observer is unstable in this version. We will fix it in the next release.

Updated on 2010-12-23. Read more

Read all announcements