If you write the raw data into a text file, you can use iconv to do a simple conversion.

Check that your iconv supports the proprietary Microsoft Arabic codepage, if you want that one, like so:

$ iconv -l | grep 1256

CP1256 MS-ARAB WINDOWS-1256

Aha, my iconv supports the proprietary Microsoft page.

So, cat YourData.txt | iconv -f CP1256 -t UTF-8 > YourData_fixed_UTF8.txt

and you should get a nice UTF-8 version (UTF-8 is the version of Unicode used mostly on the Internet).

Or, if your source was actually encoded in the proprietary Microsoft DOS Arabic codepage, 720, use -- well, my iconv doesn't support 720, so I'm not sure.

Microsoft has a beta for Arabic codepage conversion

http://www.microsoft.com/middleeast/arabicdev/beta/converter/

but it's beta, probably neither Free nor Open Source, the download does not seem to include any documentation (there is a readme which says to run the setup, which I didn't do), and my guess is it does no more (and much less) than iconv.