The Spire Script Saga
hello, i’m guster, and i’ve been trying to decipher the cipher introduced in splatoon 3’s side order DLC, dubbed “spire script”, since 15 months ago. the script is now finally on its way to being deciphered, so i wanted to detail how we got here in this post.
first of all, i did not at all do this alone; i’m part of a team of people who have all been working together for nearly a year to solve this script. the group is made up of four people: myself, cackle, @plinkamoon, and artzinil. we all bounced theories off each other, kept the group in check, and contributed ideas and tools in unique ways. magical miracle girl fia and kerfuzzle were also a huge help at many points too.
special thanks also go to: rassicas and their discord server for setting the standard for script deciphering and providing resources and a space for us to work together in; diamcreeper for providing game textures; and my friend oxii for coming up with the name “spire script”.
it’s been a hell of a journey and i’m dead serious when i say we would not have gotten to this point without every single one of these people. seriously, thank you.
the tale of spire script begins on september 14th, 2023, when the first gameplay trailer for side order dropped.
in rassicas’ private discord server, within an hour or two of the trailer’s release, artzinil noticed something that seemed like a new script on side order’s newly added official webpage.
several other graphics using the script were quickly found on the webpage, but not much could be determined from these small bits of text with little or no context.
five days later, on september 19th, 2023, graphics were released on the official splatoon twitter containing more spire script.
cackle noticed that both graphics in this tweet showed the color chip palette, and that there was script text on it—but the render of the palette had square script, while the 2D art had spire script in the same place. they checked to see if these strings matched up, and they did!
this would later be called the “palette lead”. it was the first of many potential leads that were found, and in fact the only one from the entire actual game which turned out to be correct.
cackle then tried to apply this lead to the color chips seen in the same graphic, but not too much effort was made and they moved onto deciphering things that used other, already deciphered, scripts, because this one was obviously not going to be solved using just the color chips and the text on the chips were probably gibberish.
meanwhile i try to solve the script with just the color chips
on the same day as cackle’s discovery of the palette lead, and with the exact same render from twitter, my journey with spire script began.
someone in a large discord server which i was active in at the time noticed that there seemed to be a new font in a graphic that had been posted today. i immediately took it upon myself to decipher this script with literally nothing besides the color chips (i didn’t know about any of the other places it was)
this seems like an incredibly stupid and irrational thing for me to do, and it objectively was, but it is somewhat less so when put into context. basically, i wasn’t really expecting to solve the cipher all by myself, more so just looking for something to kill time and hyperfocus on during my extremely boring classes.
my then-teammate oxii came up with the name “spire script” during practice that night, which i adopted from that point onward.
i spent hours each day for the next week or so attempting to brute-force the script using a spreadsheet text slotter (the current version has more examples than it did at the time), assuming that the text on each chip was meaningful. i also documented most potential combinations that i thought of, at first in my notes app, and then a bit later in a google doc which i would then use as a journal of sorts for all my endeavors with spire script.
on october 18th, 2023, i shared my findings about the script by adding “spire script” onto inkipedia’s page for inkling scripts and languages.
i still somehow had no idea that there was anyone else working on this script
…
and then nothing happened!
pre-release
the next pieces of spire script came five months later, with the second gameplay trailer for side order on february 13th, 2024.
i’d been sharing my findings on the script in many splatoon discord servers since september, including a modding and datamining server, and the next day, on february 14th (the 15th in their time zone), plink shared a finding in that server.
this was the 8-ball lead, based on the fact that the spire on the infinity ball seemed to match the deepsea script on octo expansion’s 8 ball (when the former was flipped around). if this lead were true, it would mean that some of spire script’s characters map to both letters and numbers.
on february 16th i finally connected with cackle on the modding server, and on february 17th i reached out to rassicas for help on tumblr and got invited to their discord server. plink joined the next day.
this is where things really started happening, now that we had the full team together and talking to each other.
the mixture of our different methods and ideas brought a lot of things about. having a text slotter spreadsheet and shorthand for all the script characters, which i’d developed back in september, were new to the members of rassicas’ server. both were very imperfect when i introduced them—the shorthand was slow to write with and i finally changed it only in december 2024, and cackle developed the text slotter into something much more organized.
the nature of spire script meant that we had to be much more thorough and resort to unconventional methods— cackle even spent hours analyzing our examples for letter frequency and comparing it to other scripts— and everyone added something to that and made it better.
the new promo material gave us several examples but only one lead besides plink’s 8-ball deduction, that being the spire script on the 8-ball goal.
it’s not a long string of characters! it’s only 4! and the words “goal” and “sink” are both 4 letters long so maybe it says one of those!
that was it that was the lead
release
the side order DLC finally came out on february 21st, 2023, introducing us to many false leads.
the most prominent of these was the pixel pearl lead, discovered by plink. there’s a pixelated pearl icon in a cutscene and when viewing dev logs— she has speech bubbles with runic script text in the cutscene, and spire script in the dev logs.
this text matched up perfectly with each other, and this method has historically been the most reliable when solving splatoon scripts. it didn’t conflict with the palette lead, either! so logically that means it’s trustworthy right??
somehow the answer is no.
this spire script text was mimicking how the runic text appeared. it’s actually evil.
pixel pearl was also used to find a few strings that seemed to be alphabet fragments, and some people used those.
there were other fake leads like this as well— the aforementioned 8-ball lead is an example, as well as text on bounce pads that lined up with the square script saying “TAKOSPO” on variants from the other story modes (found by cackle).
gas lead. i’m not sure what else to say
two leads did turn out to be correct, though, and they were found with the help of the text slotter!
two lines on the website seemed to make up words. i guessed that one of them said “name” based on what my text slotter gave me, and cackle guessed that another said “image area” from their own slotter.
these were somewhat random guesses and somehow they ended up being correct
lastly…
that’s not spire script. that’s the runic script letter for Q.
for some goddamn reason, runic Q is in spire script. our current running theory is that runic script was used as a template for spire—pixel pearl could be evidence for that—and the Q was accidentally left in and not changed.
the letter Q is one of the rarest in the english language and difficult to hit while keysmashing. so even though this is the only time runic Q appears like this with spire script, it could still be the letter Q.
does this mean we’ll have to put it on the graphic?? idk
janen ??
there were plenty other places with spire script found in side order, but none that seemed to mean anything or provide any clues. however, on july 24th, 2024, the script made a surprise reappearance in a place completely unrelated to the spire for some reason.
lemuria hub.
a 5-letter word is visible on a sign for a noodle shop in lemuria hub for some reason
this currently deciphers to “JANEN”, though any of those characters could map to a second letter. it’s probably a misspelling of “ramen”, as we already have different cipher characters for R and M. i personally think the first character could also map to D or T, since those are next to R on a keyboard, but i don’t know for sure as of writing this.
hey look it’s like blender
for some reason cackle came back to spire on august 8th, 2024, and noticed something that hadn’t been seen before.
an in-game x-y-z chart (for some 3d modeling program) has each axis labeled with a spire script character. these might actually be x, y, and z! though we don’t know which characters are which letters yet.
bancala walker
”bancala walker”, the japanese artbook for splatoon 3’s DLCs, was announced during the grand festival, on september 9th, 2024.
while the promo pages didn’t provide much in terms of a lead*, just random keymashing as usual, it did give us a preview at what would help us to begin deciphering the script.
*i did think that maybe the spire text below “drone megahorn laser” could be some corruption of those words, but i didn’t try too hard on that thankfully
the full bancala walker book released on december 13th, providing a plethora of spire strings. it didn’t go great at first…
at least i finally decided to fix my crappy shorthand.
not much happened, though
then, on january 8th, 2025, cackle’s copy of bancala walker arrived, and he found something.
he found a line of spire script that had evidently been copy-pasted from a longer string of square on the same page. this is a strategy that rassicas and company used when deciphering using examples from haikara walker, and it works just as well here, too.
it turns out that plink noting the page of every script example in the artbook on our text slotter spreadsheet, and the twitter user calmeremerald’s scan of page 19 specifically being lower quality, were what caused cackle to notice this. plink was looking at that scan for script samples and wrote down if any were too hard to make out. so when cackle got his copy, he made sure to take better pictures of the smaller script on those pages. this is what made him notice that the word lengths lined up here.
as it turns out, this text lineup from bancala walker supports, and is supported by, the palette lead (the very first lead)! it also proves the “name” and “image area” guesses to be correct.
a random cube scanner screen from the background of a stage also seems to have meaningful text. we haven’t fully figured that one out yet, though.
the next day i found another piece of matching text. this one only added one new piece of information, but it’s something!
with this, we are the most confident about spire script that we ever have been. there are, unfortunately, basically no meaningful bits of text with this script, but it was really fun to work with this team and i’m glad it’s amounted to something.