“Fancy Fonts” and using unicode characters on social media

"Fancy Fonts" and using unicode characters on social media

If you’ve been on the web for a while, you may remember the MySpace days. MySpace, preceding Facebook, was a home where you could connect typical social media style with friends, but one of the key factors of MySpace was that your profile was a little home base you could customize to your heart’s desire… To a degree. You could place all kinds of cool code to give your profile a background image, or make it play a song when someone arrived at your profile. Users could concoct really unique experiences from profile to profile.

Places like Facebook, Twitter, and others homogenized the whole experience, and with good reason; really, no one loves to have Green Day blast at them when they open someone’s profile, and then wait five hours for all the moving and flashing animated gifs to load. Plus, all the ability to customize nearly fully meant MySpace profiles could be a vector for delivering malware, spyware, and more. Allowing users a small level of customization rather than a vast all-inclusive customization experience meant it was easier to browse as a whole.

Now, in 2025, we’ve got a whole lot of homogenized social media. It can be difficult to customize your profile or stand out in a crowd of many when almost all profiles look the same, barring the profile picture, header graphic, and actual copy.

Why did we go down this quick trip down memory lane?

Because today I’m going to be talking about “fancy fonts” in use on social media, what they are, and how they affect other users.

What is this “fancy font” you’re talking about?

Sometime a few years ago, someone realized that unicode has a full set of really neat looking characters, and if you put them together, you can create something akin to a fancy looking font. For example, you can use certain unicode characters together and they look like calligraphy. Or many people overlay a lot of unicode characters to create “glitchy” looking text. Unicode is widely accepted as standard text in most social media platforms, so when one person did it, everyone started copying it. It is, after all, an easy way to bypass style limitations on social media profiles and posts, and add a little bit of flair.

These days, you can run into any sort of unicode text generator, where it will match your standard input text to a similarly paired unicode character of your choice. An example of this would be like this glitch text generator. Or this “fancy font” generator.

That’s cool! There can’t be any downsides, can there?

You may be surprised to learn why using these characters as text might be a bad idea. Let’s start by looking at what these characters are supposed to be used for, so you can understand why they exist.

Unicode characters are most often used in mathematical formulas and equations. You may also sometimes see people using other language characters to create visual emojis ( (╯°□°)╯( ┻━┻ as an example), but the ideal use case is those characters are say, inflectors for other languages. The creative use of unicode outside of their standard use case is interesting, at least, but understanding why they exist first and why simply banning/removing unicode to solve the issues I’m about to talk about can’t happen.

The main issue is such: These characters are not meant to be used to form full words or sentences. When used in such a way, visually it looks nice, but plug that into a screen reader and it’s quite the mess.

Screenshot from Mastodon instance toot.cafe from Adrian Roselli, reads "Please do not eat the shrimp scampi in the fridge. It is tasty but will kill you." using unicode characters for "do not" and "but will kill you". Post also reads "(This is an example I am using to make a recording of how it's announced by some screen readers. I will eventually reply with those videos.)" The screenshot includes a reply post with a video showcasing how a screen reader reads the post previous and says "NVDA 2024.4.2 / Firefox 135. Doesn't announce the fake bold nor fake italic."

Adrian Roselli does a great job illustrating this with a post and some videos of the output by various screen readers in this Mastodon post. One of the most egregious examples is JAWS 2025 / Chrome 133 (a very commonly used screen reader), which reads a post that visually reads as “Please do not eat the shrimp scampi in the fridge. It is tasty but will kill you.” verbally as “Please bold small d, bold small o, bold small n, bold small o, bold small t, eat the shrimp scampi in the fridge. It is tasty italic small b, italic small u, italic small t, italic small w, italic small i, italic small l, italic small l, italic small k, italic small i, italic small l, italic small l, italic small y, italic small o, italic small u.” This is not the intended effect for any user, but this is how unicode is presented to screen readers!

Even more terrifying is how the post is read by other screen readers that simply ignore the unicode characters as if they don’t exist. The post is then read as “Please eat the shrimp scampi in the fridge. It is tasty” which entirely changes the meaning of the post – no longer a warning, but instead an invitation towards certain death.

Isn’t that the screen reader’s fault? Why should I be held responsible?

You may be quick to think this is something wrong with screen readers, and shouldn’t they just be programmed differently? No, this is not an issue with the screen reader. Since we already outlined why unicode exists, it’s easier to understand with that in mind that one unicode character when presented amongst standard characters (or in the context of a formula) is performing its ideal use case, and being read as “italic small y” makes sense in that context. Screen readers cannot infer context – they just read what’s written on the screen. There is no (simple, easily available) way for a screen reader to realize unicode characters should be read as words. It is on us as posters to be aware that our words are being processed like that.

Well, I’m not a screen reader user, what else could go wrong?

Setting aside the fact that we should be working together to make the web more accessible for people with all kinds of disabilities, let’s talk about the other issues with unicode characters-as-text.

  1. Dyslexic users struggle to read many of these “fancy font” alternatives like calligraphic looking characters.
  2. Visual display output might be buggy if the platform is not prepared for this “creative” use of these; I’ve seen people on Bluesky make posts and they are blank for me, visually, because unicode was used. On Tumblr, I’ve seen unicode drip over into other posts not associated with the one using the unicode.
  3. Ⓐ looks like A, visually, but it is not. Searching for a username and/or post which uses this character requires you to use the unicode character exactly. So, you must either know the correct keystrokes on your keyboard to get it (good luck!) or you must copy-paste it in from a generator. If you want to be found via your username, plain text is the way to go.

Solutions that aren’t actually solutions

  • Shouldn’t a screen reader know how to read bold and italic fonts by now? The characters being used are not bold and italic characters, they are unicode characters, which are usually being copy-pasted in, not typed on a keyboard and formatted. Bold and italic characters are processed by a screen reader correctly.
  • Why can’t we just have Ⓐ be read as “A”? Because the actual use case for unicode means these are not the same thing. One is a standard letter, and the other is a character that holds a specific meaning for when it is used… Usually in math.
  • Can’t the screen reader just read the first instance as “italic i” and not read it out on all characters? Screen readers are not capable of inferring intent in a profile, page, post, etc. A screen reader assumes that unicode characters are being used in their proper use cases, not mis-used for visual style and flair. Unicode characters were not invented for their visual attributes, in most cases. So, if a user uses it for its initial use case after making a modification like this, then the unicode characters will be read incorrectly.
  • Damn, why you gotta be the fun police? It’s just a little stylistic flair! We really shouldn’t be torturing our disabled friends and family members just because something “looks nice.”

In conclusion…

The takeaway is this: using non-standard things to create something else is fine, but you need to understand they can have repercussions that you may not be aware of, and some of those can deal real-world harm to other people… Like eating toxic shrimp scampi. When dealing with professional accounts, it’s best to stick to the basics rather than trying to adopt something like the “fancy fonts”… And if you don’t fully know the consequences of something that seems quite mundane and silly, when someone tells you, listen to them! No one expects you to know 100% of everything, so changing your mind instead of doubling down when presented with new information is a great way to navigate the world and try to be a little kinder to our differently abled friends.

Subscribe to Foundation!

Sign up now to get notifications when new posts are made including tips, tricks, tutorials, and more. You can unsubscribe at any time! I'll never sell or share your email address.