Since creating this tool in 2020, I have read more than a million words of its output, and can thus assure you that it is the most complete, accurate, and up-to-date phonemic translator available anywhere, and it's free for all uses.
As of Jun 7, 2025, the above box can handle PDF files, but it only translates text, not images of text.
A quick visual comparison:
See how your browser and fonts render these lines
Digraph, 2025, by me. Uses digraphs for the sounds Latin doesn't have, much like the Initial_Teaching_Alphabet. The easiest phonemic script for normies to read because eight percent of words keep their traditional spellings.
Franklin, 1768, by Benjamin Franklin. Long vowels may be either doubled or circumflexed; I chose the space-saving option.
IPA, 1962, by Alfred Charles Gimson. The symbol set used by Wikipedia as well as thousands of dictionaries and ESL textbooks around the world.
CUBE, 2021, by Geoff Lindsey, who says that Gimson got the vowels all wrong, and proposes a more "evidence-based" vowel system.
SAMPA, late 1980s. A 7-bit ASCII version of IPA for QWERTY keyboards. With namer dots because capitalizing any letter changes its sound.
Diacritic, 2025, by me. Uses diacritics for the sounds Latin doesn't have, like the Czech and Vietnamese orthographies.
Shavian, 1960, by Kingsley Read. My favorite, described in great detail here. All translation to other alphabets is via Shavian, so you can copy some Shavian text from the bottom of this page, paste it into the box above, and convert it to e.g. Digraph.
Shavian54, 2022, by Evan Gallagher. Shavian plus six additional letters explained here. At present, only the Inter Alia font displays these letters correctly.
Quikscript, 1966, by Kingsley Read. A cursive form of Shavian. Download the fonts here.
Deseret, 1854. Designed by a committee, and it shows. I follow Marion Shelton's 1860 proposal of using ɪ for the unstressed schwa. Standard Deseret has no such letter; it's either omitted or replaced with various strong vowels, imposing a memorization burden on children.
Runic, 2025, by Rune Revival. This system, based on the Anglo-Saxon runes. To see bindrunes, install the Catrinity font.
Arabic, 2013, by Adnaan Mahmood. Explained here on Omniglot. Uses many variants of the letters yeh and waw to cover the English vowels.
Cyrillic transcriptions of English words are common in Russia, not just for brand names but for mundane things like "price list". Multiple phonemes map to T, A, and О, double consonants are preserved, and H can become either Г or Х, e.g. Harry Potter is Гарри Поттер but Pizza Hut is Пицца Хат.
Best of all, you can customize these alphabets or build your own from Unicode's tens of thousands of characters and combining marks. All you need is a Python interpreter, my Python program, and my American and British dictionaries.
To install the Firefox extension on your Linux computer:
###### as root
Install the DEB version of Firefox, not the SNAP version.
cd /usr/local/share
wget dechifro.org/shavian/extension.zip
unzip extension.zip
cd shavian
mv linux.json /usr/lib/mozilla/native-messaging-hosts/shavian.json
apt install python3-pip
###### as yourself
pip3 install nltk
python3
>>> import nltk
>>> nltk.download('averaged_perceptron_tagger')
>>> exit()
To install it on your Windows computer:
Install the latest version of Python. Choose "Customize installation" and check
the boxes "Install for all users" and "Add Python to environment variables".
At the Command prompt, do "pip install nltk".
In Python, do "import nltk" and "nltk.download('averaged_perceptron_tagger')"
Create a system environment variable PYTHONIOENCODING with the value "utf-8".
Download this ZIP file and extract it to C:\
Run "regedit" and add two keys:
HKEY_CURRENT_USER\Software\Mozilla\NativeMessagingHosts\shavian
HKEY_LOCAL_MACHINE\Software\Mozilla\NativeMessagingHosts\shavian
both equal to c:\shavian\windows.json
In Firefox, go to about:debugging, click "This Firefox", then
"Load Temporary Add-on", and choose the file
/usr/local/share/shavian/manifest.json or C:\shavian\manifest.json.
Every website displays normally for a second, then switches to Shavian.
This continues until you either quit Firefox or click the Extensions button
(which looks like a puzzle piece) and remove the Shavian extension.
For American pronunciation, edit glue.py line 15 to remove "brit.dict". You can also insert your own dictionary here, with changes taking effect as soon as you load or re-load a web page.
If some text appears in a low-quality bitmap font, it's because you don't have a Serif font for Shavian. Either go get one, or go to Preferences/Fonts/Advanced and set your Serif font to "Noto Sans Shavian".
Then add a function to the .bashrc file:
shave()
{
out=shavian
[ "$2" ] && out=$2
wget -N dechifro.org/shavian/{shaw.py,amer.dict.xz,brit.dict}
[ amer.dict.xz -nt amer.dict ] && xz -dkf amer.dict.xz
wget -k -U Mozilla/5.0 -O trador.html $1
echo Converting to Shavian...
( echo '<!DOCTYPE html>'
echo '<meta http-equiv="Content-Type" content="text/html; charset=UTF-8"></meta>'
tr -d @ < trador.html | python3 shaw.py amer.dict brit.dict \
| sed s/script/scrapt/g \
| sed s/twitter/twatter/g \
| sed s/NEXT_DATA/IGNORE/g \
) > /storage/emulated/0/Download/$out.html
echo All done!
}
Copy a link, open Termux, type "shave ", paste the link, press Enter, wait for it
to finish, then go to File Manager and open Download/shavian.html.
The current Shavian discussion forum seems to be Reddit, but the first was a quarterly newsletter Shaw-script wherein from 1963 to 1965 Kingsley Read received hand-written articles (in Shavian) from around the world, typed them up, and mailed photocopies to all subscribers.
I found these scans of Shaw-script hard to read owing to their peculiar font, poor image quality, and inability to search, so I trained Tesseract to read Typewriter Shavian, then proofread this 50,000-word corpus in Yudit. Here are the original text and my HTML revision of it with ligatures rejoined, end-of-line hyphenation undone, and typos corrected (e.g. Read dropped a lot of R's).
Read was grossly inconsistent in his use of namer-dots, saying that only the first word of a name needs a dot, but then dotting the next word if it fell on a new line, or if he deemed it insufficiently obvious that it was part of the name.
To remove this ambiguity, I recommend dotting all words of a name except articles and prepositions e.g. ·𐑒𐑩𐑯𐑑𐑳𐑒𐑦 ·𐑓𐑮𐑲𐑛 ·𐑗𐑦𐑒𐑦𐑯, ·𐑨𐑯 𐑝 ·𐑜𐑮𐑰𐑯 ·𐑜𐑱𐑚𐑩𐑤𐑟, ·𐑢𐑦𐑯𐑦 𐑞 ·𐑐𐑵, ·𐑻𐑕𐑩𐑤𐑩 𐑓𐑪𐑯 𐑛𐑧𐑮 ·𐑤𐑲𐑦𐑯. Aside from months, days of the week, the pronoun "I", the first word of a sentence, and the Excessively Capitalized Titles of News Articles, the namer-dot should be used exactly as capitals are in traditional orthography.
Step 0: Use Linux.
Step 1: Download a novel in PDF format
Step 2: Convert it to HTML:
pdftohtml -stdout ringworld.pdf | sed 's/ / /g' | sed 's@<br/>@ @g' | grep -v '<hr/>' | awk '{ print "<p>"$0 }' | sed s/ringworld.html//g > trador.html
Open trador.html in a web browser and make sure it looks good.
If it lacks paragraphs, try this command instead:
pdftotext -layout ringworld.pdf - | tr -d \\014\\015 | grep -v ^$ | sed 's/^ /<p>/' > trador.htmlYou won't get italics or boldface this way, but that impairs readability less than not having paragraphs.
Step 3: Run a spell-checker on trador.html. shaw.py is very good at coming up with plausible shavings for false words, which makes it a very poor spell-checker.
Step 4: Make a list of words that don't shave:
lynx -dump -width=9999 trador.html | grep -v :// | shaw.py amer.dict | tr -c '[:alpha:]' \\012 | sort -fu | grep -v '^.$' > niven.dict
Step 5: Edit niven.dict to add Shavian translations. Lowercase any words that are not proper names. The file should end up looking something like this:
$click 𐑒𐑤𐑦𐑒 $clik 𐑒𐑤𐑦𐑒 Chuft 𐑗𐑳𐑓𐑑 coffeeee 𐑒𐑪𐑓𐑰𐑰 cziltang 𐑗𐑦𐑤𐑑𐑨𐑙 Ftanss 𐑓𐑑𐑨𐑯𐑕 Halrloprillalar 𐑣𐑨𐑤𐑼𐑤𐑴𐑐𐑮𐑦𐑤𐑩𐑤𐑸 Harch 𐑣𐑸𐑗 Hroth 𐑣𐑮𐑪𐑔 Jandrova 𐑡𐑨𐑯𐑛𐑮𐑩𐑝𐑩 Kchula 𐑒𐑩𐑗𐑵𐑤𐑩 Kdapt 𐑒𐑩𐑛𐑨𐑐𐑑 kdat 𐑒𐑩𐑛𐑨𐑑 lebotom 𐑤𐑩𐑚𐑪𐑑𐑩𐑥 lookitthat 𐑤𐑫𐑒𐑦𐑑𐑞𐑨𐑑 Louis 𐑤𐑵𐑦𐑕 Rrit 𐑮𐑦𐑑 Rrrr 𐑻 sthondat 𐑕𐑔𐑪𐑯𐑛𐑨𐑑 tanj 𐑑𐑨𐑯𐑡 trinoc 𐑑𐑮𐑲𐑯𐑪𐑒 uurr 𐑻 uurrr 𐑻 wbat 𐑢𐑪𐑑 Zignamu 𐑟𐑦𐑜𐑯𐑨𐑥𐑵 Zrillir 𐑟𐑮𐑦𐑤𐑼No namer-dots, no comments, exactly two words per line. Capitalizing the first letter ensures that the Shavian word always receives a dot. "lebotom" and "wbat" don't belong here; they should have been fixed at the spell-checker step. "Louis 𐑤𐑵𐑦𐑕" is here because both audiobooks I found on YouTube say "𐑤𐑵𐑦𐑕", not "𐑤𐑵𐑦".
Step 6 (optional): Make a list of words that are always capitalized in the source text:
lynx -dump -width=9999 trador.html | uconv -x Latin-ASCII | tr -c '[:alpha:]' \\012 | sort -r | sort -ufr | grep '^[A-Z]' | sort > names shaw.py amer.dict niven.dict < names | paste -d\ names - | sed s/·//g >> niven.dictEdit niven.dict and make sure that the words you just added are all proper names and shaved to your liking. If you skip this step, some names won't receive a dot at the beginning of a sentence, and many non-English names might be shaved wrong, sometimes embarassingly wrong e.g. 𐑴𐑒𐑱𐑟𐑩𐑯𐑩 instead of ·𐑴𐑒𐑕𐑭𐑯𐑩.
Step 7: Convert the novel to Shavian:
shaw.py amer.dict niven.dict < trador.html > shavian.htmlThis leaves all heteronyms unresolved, to be fixed by hand in step 8. A part-of-speech tagger greatly reduces this chore: "shaw.py -p 1" for NLTK, "-p 2" for spaCy, or "-p 3" for Flair. NLTK is the smallest, fastest, and least accurate, while Flair is the biggest (2GB!!), slowest, and most accurate. I recommend spaCy on PCs and NLTK on smartphones and web servers.
Step 8: Edit the output file looking for @ symbols and choose the correct shaving of each one. This step does not require a Shavian keyboard.
If the text contains any references to monarchs of regnal number "I", you must find and fix them by hand, as NLTK and Flair always tag "I" as a pronoun, not a numeral.
𐑞 𐑿𐑯𐑨𐑯𐑦𐑥𐑩𐑕 𐑛𐑧𐑒𐑤𐑼𐑱𐑖𐑩𐑯 𐑝 𐑞 𐑔𐑻𐑑𐑰𐑯 𐑿𐑯𐑲𐑑𐑩𐑛 𐑕𐑑𐑱𐑑𐑕 𐑝 ·𐑩𐑥𐑧𐑮𐑦𐑒𐑩, 𐑢𐑧𐑯 𐑦𐑯 𐑞 𐑒𐑹𐑕 𐑝 𐑣𐑿𐑥𐑩𐑯 𐑦𐑝𐑧𐑯𐑑𐑕, 𐑦𐑑 𐑚𐑦𐑒𐑳𐑥𐑟 𐑯𐑧𐑕𐑩𐑕𐑺𐑦 𐑓 𐑢𐑳𐑯 𐑐𐑰𐑐𐑩𐑤 𐑑 𐑛𐑦𐑟𐑪𐑤𐑝 𐑞 𐑐𐑩𐑤𐑦𐑑𐑦𐑒𐑩𐑤 𐑚𐑨𐑯𐑛𐑟 𐑢𐑦𐑗 𐑣𐑨𐑝 𐑒𐑩𐑯𐑧𐑒𐑑𐑩𐑛 𐑞𐑧𐑥 𐑢𐑦𐑞 𐑩𐑯𐑳𐑞𐑼, 𐑯 𐑑 𐑩𐑕𐑿𐑥 𐑩𐑥𐑳𐑙 𐑞 𐑐𐑬𐑼𐑟 𐑝 𐑞 𐑻𐑔, 𐑞 𐑕𐑧𐑐𐑼𐑩𐑑 𐑯 𐑰𐑒𐑢𐑩𐑤 𐑕𐑑𐑱𐑖𐑩𐑯 𐑑 𐑢𐑦𐑗 𐑞 𐑤𐑷𐑟 𐑝 𐑯𐑱𐑗𐑼 𐑯 𐑝 𐑯𐑱𐑗𐑼'𐑟 𐑜𐑪𐑛 𐑦𐑯𐑑𐑲𐑑𐑩𐑤 𐑞𐑧𐑥, 𐑩 𐑛𐑰𐑕𐑩𐑯𐑑 𐑮𐑦𐑕𐑐𐑧𐑒𐑑 𐑑 𐑞 𐑩𐑐𐑦𐑯𐑘𐑩𐑯𐑟 𐑝 𐑥𐑨𐑯𐑒𐑲𐑯𐑛 𐑮𐑦𐑒𐑢𐑲𐑼𐑟 𐑞𐑨𐑑 𐑞𐑱 𐑖𐑫𐑛 𐑛𐑦𐑒𐑤𐑺 𐑞 𐑒𐑷𐑟𐑩𐑟 𐑢𐑦𐑗 𐑦𐑥𐑐𐑧𐑤 𐑞𐑧𐑥 𐑑 𐑞 𐑕𐑧𐑐𐑼𐑱𐑖𐑩𐑯.
𐑢𐑰 𐑣𐑴𐑤𐑛 𐑞𐑰𐑟 𐑑𐑮𐑵𐑞𐑟 𐑑 𐑚𐑰 𐑕𐑧𐑤𐑓-𐑧𐑝𐑦𐑛𐑩𐑯𐑑, 𐑞𐑨𐑑 𐑷𐑤 𐑥𐑧𐑯 𐑸 𐑒𐑮𐑦𐑱𐑑𐑩𐑛 𐑰𐑒𐑢𐑩𐑤, 𐑞𐑨𐑑 𐑞𐑱 𐑸 𐑦𐑯𐑛𐑬𐑛 𐑚𐑲 𐑞𐑺 𐑒𐑮𐑦𐑱𐑑𐑼 𐑢𐑦𐑞 𐑕𐑻𐑑𐑩𐑯 𐑳𐑯𐑱𐑤𐑾𐑯𐑩𐑚𐑩𐑤 𐑮𐑲𐑑𐑕, 𐑞𐑨𐑑 𐑩𐑥𐑳𐑙 𐑞𐑰𐑟 𐑸 𐑤𐑲𐑓, 𐑤𐑦𐑚𐑼𐑑𐑦 𐑯 𐑞 𐑐𐑼𐑕𐑿𐑑 𐑝 𐑣𐑨𐑐𐑦𐑯𐑩𐑕.--𐑞𐑨𐑑 𐑑 𐑕𐑦𐑒𐑘𐑫𐑼 𐑞𐑰𐑟 𐑮𐑲𐑑𐑕, 𐑜𐑳𐑝𐑼𐑯𐑥𐑩𐑯𐑑𐑕 𐑸 𐑦𐑯𐑕𐑑𐑦𐑑𐑿𐑑𐑩𐑛 𐑩𐑥𐑳𐑙 𐑥𐑧𐑯, 𐑛𐑦𐑮𐑲𐑝𐑦𐑙 𐑞𐑺 𐑡𐑳𐑕𐑑 𐑐𐑬𐑼𐑟 𐑓𐑮𐑪𐑥 𐑞 𐑒𐑩𐑯𐑕𐑧𐑯𐑑 𐑝 𐑞 𐑜𐑳𐑝𐑼𐑯𐑛, --𐑞𐑨𐑑 𐑢𐑧𐑯𐑧𐑝𐑼 𐑧𐑯𐑦 𐑓𐑹𐑥 𐑝 𐑜𐑳𐑝𐑼𐑯𐑥𐑩𐑯𐑑 𐑚𐑦𐑒𐑳𐑥𐑟 𐑛𐑦𐑕𐑑𐑮𐑳𐑒𐑑𐑦𐑝 𐑝 𐑞𐑰𐑟 𐑧𐑯𐑛𐑟, 𐑦𐑑 𐑦𐑟 𐑞 𐑮𐑲𐑑 𐑝 𐑞 𐑐𐑰𐑐𐑩𐑤 𐑑 𐑷𐑤𐑑𐑼 𐑹 𐑑 𐑩𐑚𐑪𐑤𐑦𐑖 𐑦𐑑, 𐑯 𐑑 𐑦𐑯𐑕𐑑𐑦𐑑𐑿𐑑 𐑯𐑿 𐑜𐑳𐑝𐑼𐑯𐑥𐑩𐑯𐑑, 𐑤𐑱𐑦𐑙 𐑦𐑑𐑕 𐑓𐑬𐑯𐑛𐑱𐑖𐑩𐑯 𐑪𐑯 𐑕𐑳𐑗 𐑐𐑮𐑦𐑯𐑕𐑩𐑐𐑩𐑤𐑟 𐑯 𐑹𐑜𐑩𐑯𐑲𐑟𐑦𐑙 𐑦𐑑𐑕 𐑐𐑬𐑼𐑟 𐑦𐑯 𐑕𐑳𐑗 𐑓𐑹𐑥, 𐑨𐑟 𐑑 𐑞𐑧𐑥 𐑖𐑨𐑤 𐑕𐑰𐑥 𐑥𐑴𐑕𐑑 𐑤𐑲𐑒𐑤𐑦 𐑑 𐑦𐑓𐑧𐑒𐑑 𐑞𐑺 𐑕𐑱𐑓𐑑𐑦 𐑯 𐑣𐑨𐑐𐑦𐑯𐑩𐑕. 𐑐𐑮𐑵𐑛𐑩𐑯𐑕, 𐑦𐑯𐑛𐑰𐑛, 𐑢𐑦𐑤 𐑛𐑦𐑒𐑑𐑱𐑑 𐑞𐑨𐑑 𐑜𐑳𐑝𐑼𐑯𐑥𐑩𐑯𐑑𐑕 𐑤𐑪𐑙 𐑦𐑕𐑑𐑨𐑚𐑤𐑦𐑖𐑑 𐑖𐑫𐑛 𐑯𐑪𐑑 𐑚𐑰 𐑗𐑱𐑯𐑡𐑛 𐑓 𐑤𐑲𐑑 𐑯 𐑑𐑮𐑨𐑯𐑟𐑾𐑯𐑑 𐑒𐑷𐑟𐑩𐑟; 𐑯 𐑩𐑒𐑹𐑛𐑦𐑙𐑤𐑦 𐑷𐑤 𐑦𐑒𐑕𐑐𐑽𐑾𐑯𐑕 𐑣𐑨𐑔 𐑖𐑴𐑯, 𐑞𐑨𐑑 𐑥𐑨𐑯𐑒𐑲𐑯𐑛 𐑸 𐑥𐑹 𐑛𐑦𐑕𐑐𐑴𐑟𐑛 𐑑 𐑕𐑳𐑓𐑼, 𐑢𐑲𐑤 𐑰𐑝𐑩𐑤𐑟 𐑸 𐑕𐑳𐑓𐑼𐑩𐑚𐑩𐑤, 𐑞𐑨𐑯 𐑑 𐑮𐑲𐑑 𐑞𐑧𐑥𐑕𐑧𐑤𐑝𐑟 𐑚𐑲 𐑩𐑚𐑪𐑤𐑦𐑖𐑦𐑙 𐑞 𐑓𐑹𐑥𐑟 𐑑 𐑢𐑦𐑗 𐑞𐑱 𐑸 𐑩𐑒𐑳𐑕𐑑𐑩𐑥𐑛. 𐑚𐑳𐑑 𐑢𐑧𐑯 𐑩 𐑤𐑪𐑙 𐑑𐑮𐑱𐑯 𐑝 𐑩𐑚𐑿𐑕𐑩𐑟 𐑯 𐑿𐑟𐑻𐑐𐑱𐑖𐑩𐑯𐑟, 𐑐𐑼𐑕𐑿𐑦𐑙 𐑦𐑯𐑝𐑺𐑾𐑚𐑤𐑦 𐑞 𐑕𐑱𐑥 𐑪𐑚𐑡𐑧𐑒𐑑 𐑦𐑝𐑦𐑯𐑕𐑩𐑟 𐑩 𐑛𐑦𐑟𐑲𐑯 𐑑 𐑮𐑦𐑛𐑿𐑕 𐑞𐑧𐑥 𐑳𐑯𐑛𐑼 𐑨𐑚𐑕𐑩𐑤𐑵𐑑 𐑛𐑧𐑕𐑐𐑩𐑑𐑦𐑟𐑩𐑥, 𐑦𐑑 𐑦𐑟 𐑞𐑺 𐑮𐑲𐑑, 𐑦𐑑 𐑦𐑟 𐑞𐑺 𐑛𐑿𐑑𐑦, 𐑑 𐑔𐑮𐑴 𐑪𐑓 𐑕𐑳𐑗 𐑜𐑳𐑝𐑼𐑯𐑥𐑩𐑯𐑑, 𐑯 𐑑 𐑐𐑮𐑩𐑝𐑲𐑛 𐑯𐑿 𐑜𐑸𐑛𐑟 𐑓 𐑞𐑺 𐑓𐑿𐑗𐑼 𐑕𐑦𐑒𐑘𐑫𐑼𐑦𐑑𐑦.--𐑕𐑳𐑗 𐑣𐑨𐑟 𐑚𐑰𐑯 𐑞 𐑐𐑱𐑖𐑩𐑯𐑑 𐑕𐑳𐑓𐑼𐑩𐑯𐑕 𐑝 𐑞𐑰𐑟 𐑒𐑪𐑤𐑩𐑯𐑦𐑟; 𐑯 𐑕𐑳𐑗 𐑦𐑟 𐑯𐑬 𐑞 𐑯𐑦𐑕𐑧𐑕𐑦𐑑𐑦 𐑢𐑦𐑗 𐑒𐑩𐑯𐑕𐑑𐑮𐑱𐑯𐑟 𐑞𐑧𐑥 𐑑 𐑷𐑤𐑑𐑼 𐑞𐑺 𐑓𐑹𐑥𐑼 𐑕𐑦𐑕𐑑𐑩𐑥𐑟 𐑝 𐑜𐑳𐑝𐑼𐑯𐑥𐑩𐑯𐑑. 𐑞 𐑣𐑦𐑕𐑑𐑼𐑦 𐑝 𐑞 𐑐𐑮𐑧𐑟𐑩𐑯𐑑 𐑒𐑦𐑙 𐑝 𐑜𐑮𐑱𐑑 ·𐑚𐑮𐑦𐑑𐑩𐑯 𐑦𐑟 𐑩 𐑣𐑦𐑕𐑑𐑼𐑦 𐑝 𐑮𐑦𐑐𐑰𐑑𐑩𐑛 𐑦𐑯𐑡𐑼𐑦𐑟 𐑯 𐑿𐑟𐑻𐑐𐑱𐑖𐑩𐑯𐑟, 𐑷𐑤 𐑣𐑨𐑝𐑦𐑙 𐑦𐑯 𐑛𐑦𐑮𐑧𐑒𐑑 𐑪𐑚𐑡𐑧𐑒𐑑 𐑞 𐑦𐑕𐑑𐑨𐑚𐑤𐑦𐑖𐑥𐑩𐑯𐑑 𐑝 𐑩𐑯 𐑨𐑚𐑕𐑩𐑤𐑵𐑑 𐑑𐑦𐑮𐑩𐑯𐑦 𐑴𐑝𐑼 𐑞𐑰𐑟 𐑕𐑑𐑱𐑑𐑕. 𐑑 𐑐𐑮𐑵𐑝 𐑞𐑦𐑕, 𐑤𐑧𐑑 𐑓𐑨𐑒𐑑𐑕 𐑚𐑰 𐑕𐑩𐑚𐑥𐑦𐑑𐑩𐑛 𐑑 𐑩 𐑒𐑨𐑯𐑛𐑦𐑛 𐑢𐑻𐑤𐑛.