I would strongly encourage anyone who knows an Arabic-script language to learn how to type in it. This isn't as difficult as one might imagine—especially on mobile devices, whose support for various languages now often exceeds that of desktop operating systems—and it opens up a whole new world of resources online. However, if you would like to use this abjad calculator but cannot type your input directly, there are a couple of alternatives. First, Google has developed clever tools for both Arabic and Persian that allow you to enter a phonetic approximation of your intended words in Roman script, and then transliterate them with very little fuss. You could use one of these tools and copy-paste the result into the box above. The second fallback option is actually a separate version of this page, in which buttons may be used to enter Arabic or Persian words one letter at a time.
By default, this program will assign an abjad value of 1 to the isolated form of hamzah (ء), treating it the same as if it were placed on an alif. Since this is not a universally agreed-upon practice, there is a checkbox to allow for the isolated hamzah to be ignored in calculation. (Please refer to the following section for more information on this dilemma and others like it.)
I've set the default letter values according to the more common "Eastern" convention. There is now an option to switch to the Maghribi order using a checkbox, for those who need it. Please refer to this table for a reminder of the differences between the two systems. (Only you can determine which of them is required by your sources.)
More broadly, the program is designed to ignore any character that it doesn't recognize. If you enter a string of text containing one or more unrecognized characters, you'll see a brief error message, followed by the total abjad value of the parts that were counted—i.e., our best guess. It may be correct nonetheless…
I recommend that you not include punctuation, or diacritics, or dagger alifs, or anything else outside of the "base letters." Only the last category is counted in the abjad system, anyway, so why add complication?
If you're confident that you entered valid text in Arabic or Persian script, but you still received an error message, please let me know by email. It might be a bug! (There are likely to be some at this stage.)
The fundamentals of abjad numerals are covered in this Wikipedia entry, and they are observed here faithfully. (I've also posted a table of letter values.) If one digs a bit deeper, however, questions and ambiguities may arise…
The treatment of the four "Persian letters"—pe, che, zhe, and gāf—is, in fact, well established. Please refer to the same Wikipedia entry. In general, there should be no problem using Persian text. (Urdu? Pashto? You're on your own for the time being.)
Actually, there is one potentially confusing point in Persian: the silent hā’. But the answer is to treat this letter as we would any other hā’. (See the following note for the reasoning behind this.) Also, if you are among the proud few who use the Unicode character 06C0 (ۀ) to represent iżāfah on a silent hā’, that will be recognized just the same.
This program treats the Arabic letter tā’ marbūṭah as if it were hā’ and assigns it a value of 5. This is due to a general principle in abjad calculation, that we judge letters by their form (rasm), rather than by their pronunciation or other factors.
Similarly, alif maqṣūrah is considered equivalent to yā’ and assigned a value of 10. (Deal with it!)
The treatment of hamzah is most troublesome of all, since it appears in conjunction with several different letters. But we need only continue to prioritize form. Hamzah on wāw (ؤ) is given the value of wāw (6). Hamzah on yā’ (ئ) is given the value of yā’ (10). Any variant of hamzah on alif, unsurprisingly, receives a value of 1. Finally, as was explained in the prior section, this program does assign a value of 1 to the isolated form of hamzah (ء), which is most often found at the end of certain Arabic words. This is a debated point; a few of the guides that I consulted are adamant that hamzah by itself should not be counted, since the appropriate value comes from the letter on which it is seated. In the end, I took the position of the Baha'i community. It has long been symbolically important in their faith that the name Bahā’ (بهاء) has an abjad value of 9, which requires that the final hamzah be counted. Users who don't want this feature may disable it with the relevant checkbox.
I should emphasize that any form of alif will be given a value of 1. This includes alif maddah, which some have treated as a "double hamzah alif " with a value of 2. Here we may invoke another general principle of abjad numerals: each letter is counted exactly once. Just as shaddah is not taken into consideration, nor should alif maddah be granted double value.