🟦 START HERE
Hello snackers and frens!
Welcome to the Filian ARG Living document! This is an ongoing informational project intended to solve the mysteries/puzzles related to the Filian ARG.
A very special thank you to our contributors & document editors for making this project what it is. And an extra special thanks to the mods, as always, for keeping communities safe and healthy.
A general explanation of how to use these Docs:
Quick Note
Hey, Just Call Me ‘L’ here (the document owner). With everything disappearing on May 1st, 2026 (twitter profile, various social banners, etc), the idea of the Snackers HQ Discord disappearing has lowkey become a real, half-hypothetical possibility.
With this in mind, I have made a dedicated server for the Filian ARG!! Note that quite a bit of things might be a little messy with the separated channels, please bear with it 😅
[when in doubt, rely on the main Filian server / “Snackers HQ” for the most up-to-date info]
The Filian ARG-dedicated Discord Server: https://discord.gg/esfqqUFBS5
Whoops some more notes about it to know before joining:
🟦 Organization Method
Generally, my (Asa) idea for this reorganization is to split everything into dedicated sections so when someone wants to know what has been tried already about anything, they can find that info (and only that info) within a few clicks. This is to avoid the choice-paralysis or info-overload of newcomers trying to parse the relevant info without having to read the entire 5 months of timeline and testing history of the whole ARG.
For ease of new players, it is recommended to mark all tabs using the following guidelines:
🟩 for solved things
🟨 for what we think might have clues that we haven't fully solved yet
🟧 for theories on the lore.
🟫 for schizo theories or very unlikely things
🟥 for already disproven things
🟦 for general info of the document
❗ next to the box where most attention and help is needed
Edit: Ok Liquid here, I got super annoyed with how disorganized this was and cleaned it up/moved things around (a second time now). Please keep theories and schizo stuff in the theories and schizo nonsense tabs. We need to keep a nice barrier between the confirmed and unconfirmed stuff. Any stuff that doesn’t fall in a specific category but isn’t a theory or schizo posting can go in the Misc Info tab. I’m going to try and keep this up to date for this whole “Phases” portion.
One last thing before you go, we (Finite Emporium on discord + whoever comes after me) have a proposed set of guidelines for the members and players of the ARG with Google doc editing permissions!
Proposed guidelines for participants with doc editing perms;
🟦 Summary of Events
Chronological order of confirmed events
Hardcore marathon:
First Unlisted Video: (Link)
@isitreallyover youtube channel: (Link)
1st livestream: (Link)
2nd livestream: (Link)
3rd livestream: (The stream is not unlisted, but rather made unavailable by YouTube since it exceeded 12 hours.)
4th livestream: (Unlisted)
YouTube Short: (Unlisted) (Low quality incomplete reupload.)
Second Unlisted (now privated) Video:
Third Unlisted Video: (Link)
Discord Server: (Invite link no longer valid)
Minecraft Server: (50.115.0.25)
The Blackfile Website: (Link)
0kind
It=stillnotfound
Sacrifice.complete=true
Initializing message...
“If I’m lost, the backup coordinates can be found on Earth near where east meets west: An ancient sea bridge, an island between B and C, bodies both. I have a yankee sixth sense about it. Two ivory towns. Carry the cross into the Mine, just the whole numbers, sixty three times raise the parallel, just twice the long.”
filians YouTube Channel: (@filianIsLost)
The Shremple:
(ONGOING?)
Current Status (Updated: May 1st 2026)
🟦 FAQs (README!!!!)
FAQs
🟦 Useful Tools
List of confirmed codes & cyphers used to solve components of the ARG so far:
Goldbug cypher
Hexadecimal code
Caesar cypher (key = 7)
Binary code
polybius cypher
A1Z26 or Modulo26
Vigenere-cipher (cryptii default key: cryptii)
Enigma-Machine (cryptii default settings: M3; rotors I, I, I; ring 1, 1, 1; pos 1, 1, 1, no plugboard; reflector UKW B)
Some Case-Specific ARG Tools: (by @hadez_mane on discord)
Toolkits:
https://github.com/gchq/CyberChef (encryption/compression analysis)
https://github.com/DominicBreuker/stego-toolkit (Steganography analysis)
https://0xrick.github.io/lists/stego/ (list of steganography tools)
https://github.com/ragibson/Steganography (Steganopgraphy manipulator)
Tools we might be interested in right now:
Audio
https://github.com/danielcardeenas/AudioStego
http://www.petitcolas.net/steganography/mp3stego/
http://www.sonicvisualiser.org/
https://www.sonicvisualiser.org/sonic-lineup/index.html [Wav comparisons between original file sample and a different sample] (Windows Startup/Shutdown noises)
https://0xrick.github.io/lists/stego/#wavsteg
https://github.com/solusipse/spectrology
https://github.com/dhsdshdhk/stegpy
Image (JPEG)
[StegSolve] http://www.caesum.com/handbook/stego.htm (Interactively transform images, view color schemes separately, ...)
[Steganabara] https://github.com/cedricbonhomme/Stegano (Interactively transform images until you find something)
https://www.boxentriq.com/steganography/stegdetect (Online StegDetect)
https://github.com/bannsec/stegoVeritas
http://old-releases.ubuntu.com/ubuntu/pool/universe/s/stegdetect
https://linux.die.net/man/1/stegbreak
https://github.com/jackfengji/f5-steganography
http://linux01.gwdg.de/~alatham/stego.html
https://github.com/lukechampine/jsteg
http://www.caesum.com/handbook/stego.htm
https://packages.debian.org/sid/utils/outguess
Text
https://www.irongeek.com/i.php?page=security/unicode-steganography-homoglyph-encoder
https://www.vincentchu.com/asciiartsteganography.html (Cursor anybody? :kekW: )
https://github.com/vgmoose/ascii-steganography
Promotional ARGs:
https://tardis.fandom.com/wiki/FindTheDoctor_(series)
(Promotional Campaign for Upcoming Series)
http://aliceetsmith.com/game/the-black-watchmen/
(ARG where one of the puzzles involved travel to an IRL Location to complete the final puzzle. Diomedes anybody? :kekW: )
🟦 Detailed Timeline of Events
This category holds the detailed info for all the confirmed events and sources that have taken place or been found in the ARG. It is ordered from oldest to newest.
🟦 Hardcore Marathon
Video Archive:
Stream 1: https://mega.nz/file/aM9G3LSZ#LTqxJ5PiMCMdAGdQ48kSN_PYUbcQff1QGCeYKhy3xHQ
Stream 3: https://mega.nz/file/XJ8TFArA#sdf_kX6C3GnSFtncS1Lqx3aujfUuWm_1iR1ct_wB08U
During filian's stream, knock sound effects and blocky visual glitches appear. After a glitch or knock, filian usually says something related to not having enough time, or having to finish on 'this run'. Note: this is not the same as the knocking bits redeem usually available, which shows when someone redeemed this, as nobody was seen redeeming these occurrences.
During visual glitches, the audio also plays a glitched out sound.
Timestamps originally compiled by Mateon1st,
redone again with links by [PHG]Tristan1999HD
Timestamps of all glitches during filian's stream (Youtube VODs) | |
Part 1 | |
2:12:17 - 2:12:18 | Glitch - “what the frick” |
3:17:17 – 3:17:18 | Knock (4) (----) - “what” |
4:33:48 – 4:33:49 | Glitch - “no! What the frick is going on?” |
5:39:21 – 5:39:23 | Knock (7) (-------) - “What the frick? What is going on?” |
6:31:34 – 6:31:36 | Glitch - “What is going on? Yo my junk’s glitching, what the frick? Hello am I good?” |
Glitch - “Not yet man. Come on, not yet” | |
Knock (4) (-..-) - “no no please no come on, not now.” | |
Glitch - “come on” | |
Part 2 | |
Knock (4) (....) - “not yet man come on, no. Gotta hurry up.” | |
Glitch (big glitch) “Chat we’re running out of time. We’re running out of time guys holy frick, no” | |
Part 3 | |
Glitch (says „a world of your imagination“ right before) - “Oh Frick, oh nonono oh god. I don’t know if we’re gonna make it” | |
3:40:33/34 | Glitch - “aw frick” |
Knock (6) (------) | |
Glitch - “aw frick, oh frick” | |
Knock (4) (-..-) | |
Glitch - “Guys we gotta win on THIS RUN” | |
Part 4 | |
glitch (big glitch) - “oh uh oh uh oh” | |
Knock (4) (....) - “ok we gotta finish this tonight, I am running out of time” | |
Glitches - “Oh no uh oh” | |
Glitch - “oh frick oh” | |
Knock (7) (-------) | |
Part 5 | |
Knock (7) (.......) - “we’re running out of time” | |
Glitch - “abelebelebele, I think this has gotta be the run, I am not long for this world” | |
User Annexes in world on join, instantly disappears (apparently Neuro’s Spectator account (Vedals alt account)?) | |
Glitch - “uh oh, yeah I really hope this is the run, would be really nice” | |
Glitch | |
Part 6 | |
Glitch - “Broo, not yet. Please a little longer x3(pleading hands)” | |
Part 7 🔴 HARDCORE MINECRAFT MARATHON 🔴Stream ends when Neuro Vedal Crelly & Filian beat the game! Pt 7 | |
Annexes joined the game (apparently Neuro’s Spectator account?) | |
Glitch (big glitch) - “you feelin’ alright? Yeah I think so, I just, I gotta hurry” | |
Knock (5) (-----) - “not yet” | |
glitch (big glitch) | |
Knock (7) (.......) | |
glitch | |
End segment | Nothing |
7:10:39 7:13:55 | Glitch - “Holy frick bro” Knock (7) (-------) |
Prerecorded clip, including the glitch documented in the ARG start tab: “Umm… (breathing in)” “Ehh…” “I didn't think it would be this soon.” “It's happening!” (unintelligible; 'worl-' maybe?) “Uhh it's.. it's.. it's time for me to go for... for a little while.” “(unintelligible; 'I'll be there in ten'/'I'm running out of time' maybe?) I really gotta go!” (out of breath) “I didn't think you'd ever see me like this!” “Ehh..” “Whuuuuh…” “You'll be able to see the world.” “My world!” | |
Timestamps in absolute time(part_time, absolute_time, seconds, type) by Makedonski
🟨 Heart Containers
Unsolved: There are heart containers that glitch across the screen during the ARG intro that no one really caught. I initially thought it was just random junk, but you can actually see it glitching into several locations at around 7:28:45 in the stream. Not sure what this asset is from. Someone said it’s a minecraft half heart but doesn’t seem to match up. Could be just an artist rendition of it.
🟩 'Time is up' link
End of Minecraft Hardcore (Part 7) 🔴 HARDCORE MINECRAFT MARATHON 🔴Stream ends when Neuro Vedal Crelly & Filian beat the game! Pt 7, Unlisted Initial Video URL was revealed for frame or two at the end:
During the minecraft marathon stream, Filian’s screen would occasionally have a glitch effect, or a knocking sound could be heard. Neither effect was triggered by chat. Whenever this happened, Filian’s expression would noticeably shift, usually followed by her saying some variation of the phrase: “I don’t have much time.”
The video URL found in the glitchy visuals at the end of the marathon stream led to the ‘time is up’ Channel via the Initial Video
🟨 'time is up' Channel
Channel → https://www.youtube.com/@isitreallyover
Channel Name → time is up
Channel Handle → @isitreallyover
Channel Created on → December 5th 2025 (users have pointed out that is the same day the minecraft marathon started on)
The channel icon for time is up when brightened in an image editing software or via display settings shows a series of dark blobs over a filian’s Twitch PFP eye. Opening the image URL directly and modifying the URL, replacing everything after = to =s0-rp gives a full size image in png format:
Brightening the full-size image reveals a string of numbers “69 20 74 68 6f 75 67 68 74 20 69 20 68 61 64 20 6d 6f 72 65 20 74 69 6d 65” which decodes via ASCII hex code to: "i thought i had more time". (Solved by fluffblush, and mypandaisnamed on discord)
Higher Quality Version (from creeperfun12_yt on Discord, originally Mateon1 in Twitch chat)
The background exactly matches the above image of filian. (credit: GeoScape in ARG discord)
Channel description (solved by Caesar+7)
vul khf... pu huvaoly spml? thfil? | one day… in another life? maybe? |
🟩 -Initial Video
Link to video (unlisted) - https://www.youtube.com/watch?v=yKYgQrogPEI
Thumbnail:
The thumbnail was found to be stock and not handcrafted for the ARG so it is unlikely to contain clues. See Transformers tweet below. Nothing was found with Aperi’Solve either.
All frames are here (processed by Makedonski) https://drive.google.com/drive/folders/1-NshvLgkTsxMMIA-rl9Nj8snmJx-WvTb?dmr=1&ec=wgc-drive-hero-goto
Video Title:
The title of the initial video is “:07 ‡0::.?*” which when put through a goldbug cipher, returns the string “YLK OLYYPUN,” which when shifted through a caesar cipher returns the phrase “red herring” which is a phrase often used to describe a clue that is meant to be a distraction from the larger puzzle. (Solved by Gold_Complex on discord)
Solution: :07 ‡0::.?* -> Gold Bug Cipher -> ``ylk ollypun`` -> Caesar Cipher -> ``Red Herring``
🟨 Livestream 1
Link to stream: 325113155552413553
Link to download: https://mega.nz/folder/hZNBnQSa#dheus_60CC24tC59tIO7Fg (Stream1.webm)
Timestamps: | |
4:52-10:20 | Binaural sinewave left=40hz right=47.83hz Rapid popping at start More popping at 5:45 Extra noises at 7:22-7:51 Popping at 8:31-10:20 |
11:47-12:37 | Crowd noise, nothing decodable? |
16:56-17:40 | phone dial (DTMF tones) (NEW) Seem to match tones at the start of VHS tapes (see The Filian ARG Document tab) |
21:30-29:30 | wind noise? |
40:36-1:11:58 | cicada sounds |
1:11:39 | KNOCKING |
1:59:46-7:10:42 | Water bubbling sounds Seems like a loop of just under 5 minutes of audio |
2:07:35 (roughly-ish) | STREAM TITLE UPDATE 325113155552413553 (Polybius: mvcezwqpx, vigenere[key: cryptii]: keepgoing) |
6:22 Audio Glitch in bubbling noise (waveform cuts out). | might have been the original hint at the 6:22:22 short |
7:11:22-7:12:52 | Beeping (heart monitor?) 49 beeps audible, a loop of 15 beeps repeated |
7:13:44-7:14:30 | Minecraft stone footstep sounds: 4,11,10,6,11,10,6,11,10,1 Then silence. Description changed, now says “its happening” encoded in binary. |
8:51:23-9:40:44 (end) | low ambiance (ominous) Wind ambiance is not a loop |
Additional Info:
Livestream thumbnail is same image as channel pfp
Video Title: Started as “.”, was eventually changed to “325113155552413553”. Using the polybius cipher gives the text “MVCEZWQPX”. Plugging the new string into a vigenere cipher with the default key on cryptii gives us “keepgoing”. (solved by ryab4ique on discord)
Solution: 325113155552413553 >> polybius cipher >> MVCEZWQPX >> vigenere cipher >> keepgoing.
Description: The stream description translates to “itshappening” in binary
Binary: | Text: |
01101001 01110100 01110011 | its |
01101000 01100001 01110000 01110000 01100101 01101110 01101001 01101110 01100111 00001010 | happening - with a new line "\n" character at the end |
Cicada noise: The cicada audio is the exact same as this Youtube clip https://www.youtube.com/watch?v=jJe9obkzf7g (found by @bloodyrose and cross referenced by @GeoScape on discord)
Stock audio for the minecraft steps was found on 2/2/26 by Random_Jawa on Discord: https://www.youtube.com/watch?v=4u8bCzEfxJM
AUDIO DOWNLOADS
Please double check the downloaded file cause someone could edit the link. Be safe!
From TongaTonk!
Water Bubbling Download (Large file 3Gb)
Low Ambiance Download (Large File 500Mb)
🟨 Livestream 2
Link to stream: jmksz o
Link to download: https://mega.nz/folder/hZNBnQSa#dheus_60CC24tC59tIO7Fg (Stream2.webm)
Livestream timestamps | |
00:53 | Music starts (“Aria Variata, BVW. 989 - Variation No. 1” by Mark Thomson - looping) |
44:10 | Audio glitch interrupts the music |
49:48 | Windows "ding" sound |
3:03:40 | Song abruptly ends |
3:05:45 | Spectrogram image of fruit snack bag, with "FILIAN IS LOST" written over it |
3:06:22 | Louder repeat of the fruitsnack image |
3:06:46 | Music continues |
3:53:44 | Music suddenly stops |
6:27:14 | Music suddenly resumes mid-note |
6:27:30 | Music pauses |
6:27:45 | Music once again resumes |
07:17:28 | Music abruptly ended once again |
11:07:06 | Stream ends (no audio playing during end) |
Video Title: jmksz o. Plugging this text into the enigma machine cypher with the default key on cryptii (all 1s) translates to resol c, or closer backwards. It should be noted that enigma machine splits text into segments of five when encoding, so resolc would be encoded to resol c.
[credits; TeleportBread for the solve and gm_flatass & liquidwater2 for theory on discord]
Solution: jmksz o >> enigma machine >> resol c >> closer
(pictured is “jmksz o” solved as “resol c” on a ‘blank’ enigma machine)
Description: The description is 114 slashes (\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\) This is as of yet unsolved/unexplained.
Exact version of the song looping:
https://www.youtube.com/watch?v=FpL8uOrPVo8&list=OLAK5uy_nt5B-j8qDzQO6SO148UwJM6Lle4FpZ4vw
Spectrogram image: (originally found by finite emporium on discord)
Direct image link (credits: Mateon1st on discord):
https://files.catbox.moe/tnhsrb.png
The font is likely Minercraftory Regular (credit to PleasDontHurtMe on Discord)
🟨 Livestream 3
Livestream was unlisted and was running for about 1h 45min before becoming public (Somebody maybe has a more accurate time stamp I was a little late so i’m guessing)
Stream timestamps | |
~1:45:00 | Stream is made public all previous video is blank, audio is fully silent |
As of 18:55:28 | Stream is still silent with no video |
As of 28:44:54 | Still all silent and without video, confirmed with ffmpeg |
32:27:58 | Suite Bergamasque, L. 75: III. Clair de lune started playing |
73:37:58 | Stream abruptly ended in the middle of music playing and video was unlisted shortly afterwards |
Additional Info:
The video is unarchived, so it is unlikely that the contents themselves are necessary for any ARG solutions.
Video Title:28 underscores (____________________________) this is as of yet unsolved/unexplained.
The number 28 might be related to BWV 28
Gottlob! nun geht das Jahr zu Ende, BWV 28 Cantata by Johann Sebastian Bach
(Praise God! Now the year comes to an end)
Another name is - ‘The Year is Ending’
This might relate to the channel “time is up” and “I thought I had more time”
Description: It was blank. There was nothing to look into.
See Clair de Lune Analysis
Stream abruptly ended and a quick check of the channel page showed the following:
Shortly after this was snipped, the video was delisted (or removed by YouTube since the livestream exceeded 12 hours).
🟩 Livestream 4
Stream Link: https://www.youtube.com/watch?v=ybVqOeumc4M (Unlisted.)
Livestream timestamps | |
00:18 | Music starts |
03:39-03:46 | Single Dial Tone beep repeated in a morse-like pattern, but starting with illegal 6 dots. ...... -.. .. -.. .. This pattern also occurs in this video. |
06:43 | Sounds like the clicking of opening a safe, interrupted by videogamey "failure" sound (turned out to be blackfile-index.org sounds) |
08:55 | Dial Tone beeps (exact repeat) |
09:30 09:50 | Clicking sounds and "failure", single repeat each |
14:11 | Dial Tone beeps |
19:27 | Dial tone beeps |
~33:00 | Three more sections of clicks and “failures” |
At 12:40 pacific description changed for about 10 seconds saying the following in the middle of the slashes: | N̸̯̭͒͠o̴̳͉̚̕w̸̻̏ ̸̧̠͛̊t̸̪͝h̶̹̞̏a̵̭̔͜t̴͉̲̿̚ ̸̭͍͌́y̷̧̏͗o̸͙͒u̸̖͆ ̷̧̞̀h̸̰̑̆ȧ̷̯͎v̷̪̀ḙ̵̺̚ ̶̬̄̑ḛ̵̐͝v̴͎̟͝͝ė̸̪r̷͎̀ȳ̵̙̥t̶̩̼̀h̶̙͐i̷̳̊n̶̩̜̄͝ġ̸̡̍.̸͍̎ ̵͙̋ k̶̘̎ẹ̷̋e̶̸̞̯̯̬̓̊p̵͍̋ ̴̮̔ͅl̷͛͜͝o̵͉͂o̴̡̅ḱ̸̝͖͝í̶̛͖n̸̖̿g̴̝͐ |
12:45-50~ | Title was changed to “.” aka a single period |
2:02 AM pst | The following unlisted youtube link was posted in the chat by the channel account. Stream ended about 2 seconds later https://www.youtube.com/watch?v=qJ98vSrX 2uU |
Additional Info:
This stream had static frames rather than a pitch black screen. In this way it was different than the previous three. More info on the static frames tab.
For those who are struggling to read the garbled text, it says
Now that you have everything - keep looking
Music: Sonata in G Minor, Op50 No.3 - Didone abbandonata (Clementi), original per Discord user Skragger: https://www.youtube.com/watch?v=zxY0HHos_9U
https://en.wikipedia.org/wiki/Didone_abbandonata_(Clementi)
"Didone abbandonata" translates to "Dido Abandoned" in English. It refers to the tragic story of Dido, the queen of Carthage, who is abandoned by her lover, the Trojan hero Aeneas, as he departs to fulfill his destiny of founding Rome. The title is derived from the myth recounted in Virgil's Aeneid, where Dido, heartbroken by Aeneas' departure, ultimately takes her own life by throwing herself onto a funeral pyre.
Morse code sounds and the clicking sounds line up. Credit: Liquid (@liquidwater2 on discord)
🟩 Final Video/link
At about 2:02 AM PST, an unlisted video was linked in the 4th Livestream chat:
https://www.youtube.com/watch?v=qJ98vSrX2uU
Title: / /
Description: (no description)
This video was removed shortly after and is now unavailable as it was potentially made with errors, 6 2 2 instead of 6 22 22
At about 2:15 AM PST, another video was posted on the ARG channel’s Posts tab.
At about 2:15 AM PST, another video was posted on the ARG channel’s Posts tab (https://www.youtube.com/@isitreallyover/posts):
https://www.youtube.com/watch?v=z0DeGdzKZGg
Title: / /
Description: //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////_/__/__//////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
🟩 6/22/22 Analysis
At 5:01 AM, a discord server link with corrupted text was found in the description of a short posted on 6/22/22, and was considered to be the solution to livestream 4.
(Credit: liquidwater2 for discovering the code was 6 22 22 and referencing a date/time, antuber_46672 for finding the short posted on that date, and Dababe31 for finding the discord invite link)
Short Description:
Discord server link: https://discord.gg/ZJ6aewaP
Honorable Mention:
In the last twitch livestream (the one that includes the start of the ARG at the end), Filian dies at exactly 6:22:22.
🟨 ARG Discord Server
The original invite to the server has expired.
Please see the contained tabs for our analysis.
Server name: //////////////////////// (24 slashes)
Additional Info:
If you grab the discord server owner’s icon and brighten it you seem to get the exact same picture as the time is up channel icon with the hexcode for ‘I thought I had more time’.
🟨 Owner Account Analysis
Created by user: _._.__.___.____.___.__._._
Update: 12:20 UTC Dec 23, 2025. Makedonski found confirmation that the numbers at the end of the owner’s username were automatically added by discord. Source
Additional important information from discord username page:
1. Usernames cannot use 2 consecutive period characters ( . )
2. Usernames cannot use any other special characters besides underscore ( _ ) and period ( . )
Pythonhare suggested that the dots are actually meant to be slashes:
_/_/__/___/____/___/__/_/_
Original screenshot of the discord / owner’s info | Most recent screenshot after owner’s username was changed automatically by discord (screenshot credit The Midnight Station / themidnightstation. on the Snacker’s HQ discord) |
🟩 Spectrogram
Voice channels were made available for a short time with names ".." ".." ". ." ". . ." and the last one allowed some users to play a soundboard.
The soundboard in the discord server had an audio file which showed the IP address for the Minecraft server as an image on the spectrogram:
This spectrogram with the IP was found at 10:54pm PST on Dec. 19th, 2025 and a Minecraft server with that IP was connected to a few minutes later by 11:01pm PST.
🟨 Minecraft
Minecraft Server
Gamemode: Hardcore - if a player dies, they can only spectate afterwards
Title: Originally “find it”, temporarily changed to “If only you knew how much more is out there” after the second structure was found, then almost instantly changed to “find. it.”.
Version: Java 1.21.11
Seed: “711026799995424918”
Server is hosted on apexhosting with a multicraft console (GeoScape on Discord):
Initial Finds:
Command Block Coords:
1123, 90, -1853
1123, 89, -1853
Timeline:
🟩 Revive Altar
Photos of the altar and signs:
Location is -401, 101, -3075 roughly
Solution:
Signs are originally in the galactic alphabet (enchanting table alphabet), translated they say the following: (thanks corrizer on discord)
If you want your fallen a sacrifice must be made
Place the head of the dragon at the top
Dont misunderstand theres more to be found
🟩 Filian Temple
Photos of the temple structure:
🟨 Shremple
THE SHREMPLE
Candle Info | Corner Text Ordering Possibilities | ||
Candle Number | Candle Color | Corner text (interior glass) → Also placement order in time | Corner text (candle location) |
1 | blue | remember. | protect it. |
2 | white | watch it | its coming |
3 | purple | protect it. | remember. |
4 | black | its coming | watch it |
Details of the shremple layout (by Asa) (small additions by EuphoricFan1161)
(This is just every single thing I noticed as I wandered around the temple, there is probably more I didn’t notice.)
You can check out an unmodified version of the Shremple by following these instructions: Partial World Download
Under the Temple:
Black pillar of concrete down to bedrock in only one corner under temple (Purple 3-candle side / NorthEast side)
- includes chiseled quartz block on the corner - no other corner has this
Towers on Temple:
Pillars on top of temple are Quartz Pillar which stop at the floor where they meet
- Between the 4 towers are two mirrored ornate boxes. On the South side, pink glass fills the window/cavity but nothing on the North side
- The stair pattern in the wall of the towers is asymmetric about the x-axis but symmetric about the z-axis
- Stair-step level of temple floor causes Quartz Pillar to stop 3 blocks higher on inside of temple than outside of temple
- Top 'Trim' of Pillars hides a row of Quartz Pillars
- Top “trim” of the southeast pillar (the remember pillar) is missing its quartz slabs in the corners
- Quartz Pillars go down to bedrock 4 blocks below floor of temple
- Quartz Pillars contain unique seemingly-random patterns of Chiseled Quartz, Purple Glass, and Black Glass
- Quartz Pillars have a colored candle next to each, with a different number of candles in the colored stack: Blue=1, White=2, Purple=3, Black=4
- Quartz Pillars are each labeled with their own floating text: ‘protect it.’, ‘its coming’, ‘remember’, ‘watch it’
- Only the ‘protect it.’ and ‘remember.’ floating text labels have periods while the other two do not.
Inside Temple:
- Floor has cavities in corners of a 7x9 rectangle, with 2-tall Iron Pyramids below (filled with Beacons immediately on the live server)
- 3x3 rings of blocks of colored glass matching candle colors, empty middles (except one)
- The Black 3x3 ring is the only one with a filled center, which is Chiseled Quartz
- Single Wither Rose between each pair of 3x3 rings
Missing Stairs:
- On the southeast side, both chiseled block rows above the 0th tier are missing their terminating stairs
- Near the northeast corner, on the east side, the chiseled block row above the 3rd tier is missing its terminating stairs
Chiseled Blocks:
- Northeast side has chiseled quartz block on the lowest corner before the black concrete
- Under the north side of the eastern stairs there is a chiseled quartz block one block lower than the corner chiseled quartz block
- Next to the southwest pillar on the west side of the 2nd tier (starting count from 0), the lowest block is a chiseled quartz block
- Also on the 2nd tier on the northwest side, 9 blocks south of the pillar and directly behind the chiseled quartz block row is another chiseled block
- Interior, near each set of 2 3x3 glass rings, the chiseled blocks are not mirrored. The eastern side has 3 shaped in a small L, the western side has 3 in a vertical line. The cavities behind are similarly asymmetric about the x-axis and the western side is asymmetric about the z-axis as well
- The small towers contain many asymmetries about the z-axis in chiseled block positions
We have not found the next step after finding the Shremple.
🟦 Checked Coords
A few notes on the limitations of our scanning efforts:
- Spectators are unable to load chunks past 31700 blocks unless those weren’t generated by other alive players (This in theory would include the GM when they placed the Shremple before the ARG started, so it’s unclear why those chunks were not loaded during our scan)
- PC__ on discord scanned the whole world with esp, however I (Asa) have been unable to determine the exact method by which this was done and/or why it missed some things such as the Shremple. - PC here, I scanned all of the overworld manually, not automated, from March 1st-28th. Died on purpose to search as a spectator, used meteor esp (necessary evil for this scale) Whitelist too long for here. Shremple is just outside the spectator range on the right side which is why I didn’t bump into it before the riddle. I encourage anyone to search the Nether.
We have performed automated scans on the world using the following methods:
- Metakimi on discord scanned at least half of the Overworld with a scan tool Darrien made and 10k in every direction in both the Nether and The End.
- Pizzabash on discord scanned 15k in the nether.
- Liquid on discord covered The End area surface way before the server had (the?) maintenance pause(s?).
Additionally, we have several tables of coords that have been checked by hand.
These lists are periodically backed up but editing is open.
Minecraft Coords.xlsx [Owned by Darrien 🙏🙏]
checked coordinates [Maintained by Not_Here]
Please keep notes on coordinates, thorough checking of possible locations will reduce need for rechecking.
Additionally, these are even more coords that have been checked:
Checked coords by Stardrop:
elytra:
7374 138 4914
5599 90 6195
5196 112 -5778
4791 119 4227
4630 104 -4732
4588 123 -3710
4174 137 5193
4375 138 5914
4197 108 -4048
4143 112 -5073
4277 139 -5456
3960 79 3585
3039 135 1981
2719 111 4236
2703 107 1668
1617 30 439
1559 115 3919
1282 128 -4449
1116 96 5121
606 99 -2802
673 112 -4493e
-17 127 4859
-278 107 4643
-968 109 5234
-897 97 4335
-1201 85 4868
-1181 91 1423
-2149 92 613
-2127 130 -146
-2563 134 4546
-2939 126 85
-3129 103 -338
-3567 131 -1596
-3429 140 4222
-3689 115 3274
-3845 135 -1534
-3878 132 -2821
-3818 103 -4730
-3859 124 404
-4141 80 -2519
-4806 138 4187
-5129 124 -3743
-5087 83 2899
-5120 82 3883
Nether :
1123, 97, -1853, (try nether portal here later)
1617 30 439 (big lava lake, maybe try with a remove lava plugin but didn’t find anything under it)
Checked coords Nether and End (all Y):
X = 181, Z = 37
X = 358, Z = 37
X = 181, Z = 38
X = 358, Z = 38
X = -2, Z = 88
X = 0, Z = 88
X = 1, Z = 88
X = 2, Z = 88
X = 181, Z = -1575
X = 358, Z = -1575
X = 181, Z = -1635
X = 181, Z = -1638
X = 358, Z = -1635
X = 358, Z = -1638
X = -2, Z = 1575
X = 0, Z = 1575
X = 1, Z = 1575
X = 2, Z = 1575
X = -2, Z = 1634
X = 0, Z = 1634
X = 1, Z = 1634
X = 94, Z = 11574
X = 144, Z = 7446
X = 941, Z = 1574
X = 1122, Z = 4646
X = 1447, Z = 446
X = 9411, Z = 574
X = 14790, Z = 320
X = 32041, Z = 38
X = 32041, Z = 37
X = 32143, Z = 37
X = 32041, Z = -1575
X = 32041, Z = -1638
X = 32143, Z = -1635
7600, 16385
6002, 18005
end gateway:
SPAWN -> 992 73 343
5000,5000
-5000,-5000
4951 68 5124
5318 70 1794 (two)
5317 73 -4339
5326 69 -4979
4838 66 4943
4584 70 -3781
4559 62 4120
4591 69 5123
4352 72 1791
4315 75 1195
4196 68 -295
4209 67 -3663
4013 70 -4203
4015 69 -222
4013 70 -222
3893 68 -5479
3659 67 -4811
3731 73 1167
3769 67 2448
3728 72 3249
3542 85 4909
3429 66 1979
3356 130 -861
2928 72 3937
2957 69 4732
2739 72 3471
2743 69 1735
2604 71 -2739
2401 71 -5224
2307 66 -1809
2422 65 -966
2442 71 1935
2350 62 2506
2517 63 4120
1951 69 379
1721 69 -4867
1666 67 -3515
1798 67 -2459
1554 66 4777
1406 71 3781
1273 73 -4381
1173 65 -4962
941 66 -3283
987 71 -2825
915 66 3448
990 64 4238
861 67 4271
784 69 2426
718 66 -2784
305 70 -3280
162 71 -1395
323 66 4653
-175 73 -2328
-240 58 2551
-590 68 -3475
-529 72 4238
-785 71 -3003
-1009 70 -2580
-899 75 -1205
-1045 68 5315
-1220 71 5121
-1322 71 -603
-1655 69 2376
-1989 71 1418
-1974 66 -2557
-2103 67 -3426
-1994 74 -3516
-2128 64 1033
-2386 69 1714
-2728 72 2603
-2714 69 -1051
-2695 71 -2998
-2760 66 -3850
-3042 66 -2846
-2985 67 -2765
-2959 69 3412
-2901 74 4055
-2889 70 4719
-3127 67 4268
-3182 65 3044
-3155 67 739
-3202 66 6766
-3086 91 -5317
-3403 74 -4789
-3579 68 5964
-4280 70 -1216
-4145 75 -279
-4593 70 3900
-4867 67 6056
-4751 65 377
-4702 70 -1329
-5064 73 -2123
-5054 70 -116
Checked coords Overworld:
65 -168 (Bedrock to Build Limit)
128 -166 (Bedrock to Build Limit)
128 -170 (Bedrock to Build Limit)
7600, 16385
6002, 18005
List of tried coordinates so far (not up to date)
note: Minecraft has flipped coordinates from expected → x=longitude, z=latitude
-z is north and +z is south
real conversion is 1 degree = ~111 km or ~111,111 blocks hopefully not this
Overworld (x/[y]/z)
Nether (x/[y]/z)
End (x/[y]/z)
Coords are accurate to Stronghold portal room spawner block (nether roof ports y = 128)
All end portals should now be activated, none appeared activated without intrusion
X Y Z Active? Notes
🟦 Partial World Download
https://drive.google.com/file/d/1MAlw_vjH7qNvgoDRntkGIQs06DIf13dr/view?usp=sharing
The link is to a drive with a world file of the temple untouched before discovery
To use
Computer Steps
Minecraft Steps
Computer Steps 2.0
Minecraft Steps 2.0
World download courtesy of aezukla on Discord give them many headpats if seen
🟨 blackfile-index.org
Site is currently down. No progress can be made here.
Website Link: https://www.blackfile-index.org/
See this tab for how this website was found: Filian Temple
See this tab for info on the windows 95 desktop environment: WIN95 UPDATE
Corrupted file error
Error Reading Drive error
Access Denied error
Incorrect file password error
Desktop accessible tabs:
As of March 24th the open mysteries with this website are the following:
Broken cursor
On both css’s for the terminal and for the desktop, there is a cursor object that does not render. Many attempts have been made at visualizing it or deciphering it. It’s unsure whether or not this cursor is a red herring or has something encoded in it. Additional info can be found here: broken cursor
Text Version of convo:
Me: hey.
J: Hey. Any updates on the girl?
Me: Not yet…
J: “...” ?
Me: There’s been a hiccup.
J: What do you mean? I don’t like how vague this sounds.
Me: It looks like anonymous users have been interacting with the
desktop environment…
J: How? Everything’s encrypted.
Me: Who knows. Hackers? Shoddy programming? Internal leaks? Hard
to say.
someone could be feeding the public info as well…
J: Have they viewed any sensitive info?
Me: no. The docs are still password protected, nobody’s found or shared the pw to anything there.
J: Alright. Keep it that way.
🟩 Password Terminal
Terminal Passwords:
🟨 README.md
Read Me Page:
A readme file was discovered at https://www.blackfile-index.org/README.md and believed to have been accidentally left in. However, the text began counting down daily starting on 12/30/2025
Day 0: # 4kind-arg
Day 1: # 3kind arg///it=notfound
Day 2: # 2kind arg///it=notfound
Day 3: # 1kind arg///it=notfound
Day 4: # 0kind arg///it=notfound
One day after the countdown reached 0 this message appeared:
0kind
It=stillnotfound
Sacrifice.complete=true
Initializing message...
“If I’m lost, the backup coordinates can be found on Earth near where east meets west: An ancient sea bridge, an island between B and C, bodies both. I have a yankee sixth sense about it. Two ivory towns. Carry the cross into the Mine, just the whole numbers, sixty three times raise the parallel, just twice the long.”
The consensus is that the most likely explanation is that this is pointing to an island in the Bering Straight (either Little Diomede or St. Lawrence) and saying that we need to convert the coordinates of that island into minecraft.
There have been attempts to reverse engineer the answer using the shremple coordinates, this can be found here
This was never solved and many think it was replaced by the konami code riddle.
🟩 -Konami Code
Konami Code
Timeline:
The five notes played at the beginning of the message are C, B, A, C, G (all in the 4th octave). When each note is converted to its position in the alphabet, this gives 3, 2, 1, 3, 7, forming the number 32137, so 32137 is X.
Water is technically filtered by boiling. The boiling temperature of water is 100° Celsius. It also says lose 127, so doing 100 - 127 yields -27, so -27 is Y.
All the mentioned events from the last segment date back to the 1700s. The capitalization spells out the word CENTURY. 1700s are the 18th century, so 18 is z.
🟦 Code Deobfuscated
Here is a repo of all of the data on the source code of the blackfile-index website that we have:
https://drive.google.com/drive/folders/181gUaxFIuxnk0leSdHnM2FnUw9-RkgnO
It includes the payload injection, the obfuscated code, and the deobfuscated code for every minor update/change made to the website by the GM.
🟦 Checked Passwords
We have a bot brute forcing the password, not that we really think this is the intended solution.
Extreme Caution is required if you’re thinking about creating your own brute forcer. It is VERY easy for this kind of method to accidentally turn into actual illegal activity.
Generally, ARG passwords are intended to be solved by completing the puzzles rather than simply checking every possible combination of chars. We are not completely confident brute forcing is ‘in-bounds’ of the ARG.
The password will be (or likely already has been) given to us somehow. It’s just a matter of finding it.
By makedonski (sqlite database, around 11.5 million records, updated daily) : codes
Additionally, these passwords have been manually checked:
–TODO - add list of passwords derived from potential puzzle solutions
🟦 Dokomi Con / Advert
On the official Dokomi.expo livestream, Filian can be heard speaking just before the QR Code is shown. The QR Code links to a filianislost website.
Link to advert re-upload:
https://youtu.be/1zVfLq0TWG8
These are the gifs used in the video background. (found by liquid)
Filian’s voice appears in a purple audio visualizer which is overlaid on these gifs.
Filian’s voice lines are the following:
Hey guys!
Wha-
Wha-
Wait up!
Wo willst du-
Wo willst du hin??
No.
Wo willst du hin??
Wo willst du hin??
Where do you think you're going!?
I have a-
❗ filianislost.com
The filianislost.com website was created on 2/12/26
On 5/29/26 at around 3:34 am UTC, snackers noticed the website filianislost.com changed from the text: “HAVE YOU SEEN HER” to a countdown ending at her normal stream time on june 1st labeled “Phase 1”, and a box to enter your email to “get notified”.
Swiika has retweeted the website five times, so it is very likely to be legitimate:
🟦 THE_WALL
THE_WALL
PHASE 1 Timeline:
June 1st:
The countdown on the filianislost.com website ended, and it became a painting canvas over the top of the following stock image of a wall. It included a section to submit “signals” which seem to be grassroots style advertisements for filian.
June 5th:
An email was sent out to the email list with the following information:
“The Wall Is Yours To Lose. Five days in and the wars have already started. People are claiming territory, painting over each other, and leaving their mark for good. This weekend, paint refreshes twice as fast. Double Paint Weekend is here. If you haven’t shown up yet, someone already has your spot.”
“TERRITORY WARS ARE REAL. People are claiming space. Others are painting over it. Alliances are forming. If you haven’t staked your claim yet, someone already has. Go protect your territory or take someone else's.
This Saturday only from 6PM to 10PM EST - when your refresh timer runs out you get one extra paint. Double the paint. Double the territory. If there was ever a time to make your move it's now.”
“LEADERBOARD REWARD. The top 10 painters on the leaderboard will each receive an exclusive award poster of the Wall. We’ll reach out via email. Current top 10 you know who you are. Everyone else, you have a weekend to catch up. Bring her back.”
The double paint event did not take place until 3 weeks later on the 27th, 28th, and 29th.
June 10th:
A poll took place to wipe a quarter of the wall. The bottom right corner was wiped.
June 17th:
We were informed via @THE_W4LL_ twitter account (listed below) that the final result of the wall would be displayed on a billboard at shibuya crossing at the end of phase 2, and that phase 2 would be a continuation of the wall beginning on June 19th and last until June 27th at 11pm central.
https://x.com/THE_W4LL_/status/2067433113152573912?s=20
PHASE 2 Timeline:
June 19th:
The wall flashed a lost signal tv screen overlay, then was wiped along with both leaderboards. A countdown was added to the website for the end of phase 2.
June 22nd:
A video taken in shibuya crossing showed an ad for the filianislost website on one of the billboards https://x.com/RimaEvenstar/status/2069079309302497605?s=20 The qr code and link are to the filianislost website. A few different angles of the ad have been posted at different times of day by various reddit and x users. A snacker living in japan went to the billboard to see if they could confirm the legitimacy of the ad and it never played on the billboard while they were in shibuya crossing.
Liquid noticed that the horizontal glitching effects on the video are actually lines of text that say Filian is lost/missing in different languages as shown in the following images:
Higher quality of the billboard is needed to see if any other phrases exist.
June 23rd:
Another email was shared with the following text
“Phase 2 - Shibuya
Your Art. Shibuya Crossing. When the timer hits zero, whatever is on The Wall goes up on the billboard at Shibuya Crossing. One of the most iconic screens in Tokyo. Every pixel you place right now is part of what the world sees.
This is what’s at stake. Millions of people pass through Shibuya Crossing every day. Your mark on the wall could be one of them. Claim your territory. Protect it. Or take someone else’s. Either way, make it count.
Don’t let someone paint over your spot. The canvas is live right now. People are painting, claiming territory, and leaving their mark. The timer is running.”
June 26th
Two days were added to the timer at 7:16 A.M Central, making the new ending June 29th at 11pm Central No statements or announcements or have been as of this update, the timer simply changed on the website.
Later the same day the following notification was sent out to users on the website.
—------------------------------------------------------------------------------------------------------------------
The emails above included links to the wall’s twitter page:https://x.com/THE_W4LL_
And the wall subreddit: https://www.reddit.com/r/THE_WALL_/
Most videos on the X account have glitch effects and/or CRT screen effects. Many seem to be made in an unsettling old tv commercial style. These have been frame by framed and nothing was found.
June 29th:
The phase 2 countdown ended and changes can no longer be made to the wall. The following message now appears when you open the site;
The wall’s X account informed us that the top 10 from the end of phase 1 will be reached out to regarding the posters. As of June 30th phase 3 has not yet been announced, but via the above image the wall seems like it will be re-opening.
July 7th:
At 6:07 PST the message on the wall saying it is closed was removed, and users are again able to paint on the wall. No announcement has been made about this on x or the site itself.
THIS WILL BE UPDATED AS NEW INFO EMERGES
🟦 Cursor Info
Cursor Names
How to see the animated cursor: go to https://filianislost.com/static/cursors/generated/[number].png and replace [number] with the cursor number you want to see
As of June 2nd at 2:51Pm the total amount of cans needed to unlock every cursor with coins is 189. For painting purposes it requires 63 cans to unlock all current pixel related cursors. The grand total being 252 cans to unlock the currently available cursors. With a 5 minute cooldown it will take roughly 21 hours worth of cans to unlock everything.
Cursor | Cursor | Name | Dimensions | Animated | How to Get | What set of cursor releases did it become available (WIP) |
0 | Default #0: White | 32x32 | No | Default | Group 1 | |
1 | Default #7: Blue | 32x32 | No | In inventory by default | Group 1 | |
2 | Default #1: Dark | 32x32 | No | In inventory by default | Group 1 | |
3 | Default #4: Yellow | 32x32 | No | In inventory by default | Group 1 | |
4 | Default #5: Green | 32x32 | No | In inventory by default | Group 1 | |
5 | Default #3: Orange | 32x32 | No | In inventory by default | Group 1 | |
6 | Default #9: Pink | 32x32 | No | In inventory by default | Group 1 | |
7 | Default #8: Purple | 32x32 | No | In inventory by default | Group 1 | |
8 | Default #2: Red | 32x32 | No | In inventory by default | Group 1 | |
9 | Default #6: Sky Blue | 32x32 | No | In inventory by default | Group 1 | |
10 | Asexual Pride | 24x24 | No | Purchase (20 coins) | ||
11 | Bisexual Pride | 24x24 | No | Purchase (20 coins) | ||
12 | Blue Butterfly | 32x32 | Yes | Purchase (15 coins) | ||
13 | Christmas Tree | 32x32 | Yes | Purchase (10 coins) | ||
14 | Diamond Sword | 32x32 | No | Purchase (15 coins) | ||
15 | Fireblaze | 32x32 | No | Purchase (10 coins) | ||
16 | Gay Pride | 24x24 | No | Purchase (20 coins) | ||
17 | Genderfluid Pride | 24x24 | No | Purchase (20 coins) | ||
18 | Golden Arrow | 32x32 | No | Purchase (15 coins) | ||
19 | Golden | 32x32 | Yes | Purchase (10 coins) | ||
20 | Lesbian Pride | 24x24 | No | Purchase (20 coins) | ||
21 | Nonbinary Pride | 24x24 | No | Purchase (20 coins) | ||
22 | Pansexual Pride | 24x24 | No | Purchase (20 coins) | ||
23 | Party Hat 2 | 32x32 | No | Purchase (10 coins) | ||
24 | Party Hat 3 | 32x32 | No | Purchase (10 coins) | ||
25 | Party Hat 4 | 32x32 | No | Purchase (10 coins) | ||
26 | Party Hat | 32x32 | No | Purchase (10 coins) | ||
27 (T1) | Phatblue | 46x48 | No | “This is a mod only cursor” | Group 1 | |
28 | Pink Cursor | 48x48 | No | Purchase (15 coins) | ||
29 | Pulsing Blue | 32x32 | Yes | Purchase (15 coins) | ||
30 | Spinning Circle | 32x32 | Yes | Purchase (10 coins) | ||
31 | Transgender Pride | 24x24 | No | Purchase (20 coins) | ||
32 | WinXP | 32x32 | No | Purchase (10 coins) | ||
33 (T2) | 420 | 32x32 | Yes | |||
34 (T2) | AK-47 | 32x32 | Yes | Purchase (47 coins) | ||
35 (T2) | America Cursor | 32x32 | No | Purchase (47 coins) | ||
36 | Bluegem Blue | 32x32 | Yes | Purchase (40 coins) | ||
37 | Bluegem Green | 32x32 | Yes | Purchase (40 coins) | ||
38 | Bluegem | 32x32 | Yes | Purchase (40 coins) | ||
39 | Chainsaw | 32x32 | Yes | Purchase (67 coins) | ||
40 | Chroma | 32x32 | Yes | Purchase (40 coins) | ||
41 | Bone | 32x32 | No | Purchase (25 coins) | ||
42 | Dinosaur | 32x32 | No | Purchase (21 coins) | ||
43 (T2) | Yoda Lightsaber | 32x32 | No | Purchase (40 coins) | ||
44 | Chihuahua | 32x32 | No | Purchase (21 coins) | ||
45 | Earth | 32x32 | Yes | Purchase (40 coins) | ||
46 | EXE 1 | 32x32 | No | Invite 1 member to the board (condition likely changed, received w/o doing this) | ||
47 (T3) | Flipping the Bird | 32x32 | No | |||
48 (T3) | Gold Orb | 32x32 | Yes | |||
49 | Labradoodle | 32x32 | No | Purchase (21 coins) | ||
50 (T3) | Lineage 2 Sword | 32x32 | No | |||
51 (T3) | Lineage 2 Arm | 32x32 | No | |||
52 | Pink Heart | 32x32 | No | Purchase (25 coins) | ||
53 | Pug | 32x32 | No | Purchase (21 coins) | ||
54 | Konami Cursor | 32x32 | Yes | Input the Konami code | Group 1 | |
55 (T3) | Swords - Claim Solais | 32x32 | No | |||
56 (T3) | Swords - Elements | 32x32 | No | |||
57 (T3) | Unicorn | 32x32 | Yes | |||
58 | Wand | 32x32 | Yes | Purchase (10 coins) | ||
59 (T3) | WIB | 32x32 | No | |||
60 | Working in Background | 32x32 | Yes | Paint 100000 pixels. Each can holds 1000 pixels so you need to empty your can 100 times | ||
61 (T4) | Support Bird | 32x32 | No | Invite 5 people | Group 1 | |
62 (T4) | Blue Jack-o-Lantern | 32x32 | No | |||
63 (T4) | Dark Jack-o-Lantern | 32x32 | No | |||
64 | Ghost Face | 32x32 | No | Purchase (67 coins) | ||
65 (T4) | Gun | 96x93 | No | Purchase (100 coins) | ||
66 | Ovaley | 32x32 | Yes | Purchase (25 coins) | ||
67 (T4) | Pink Jack-o-Lantern | 32x32 | No | |||
68 | can i haz cursor plz | 32x32 | Yes | Type “can i haz cursor plz” in chat | Group 1 | |
69 | Gun 2 | 96x90 | No | Purchase (67 coins) | ||
70 (T4) | Gun 3 | 128x128 | No | Purchase (67 coins) | ||
71 | Tutorial Bird | 32x32 | No | Read through the guidebook | Group 1 | |
72 | Flame Black | 32x32 | Yes | Purchase (30 coins) | ||
73 | Flame Aqua | 32x32 | Yes | Purchase (30 coins) | ||
74 | Flame Black Blue | 32x32 | Yes | Purchase (30 coins) | ||
75 | Flame Black Red | 32x32 | Yes | Purchase (30 coins) | ||
76 | Flame Blue | 32x32 | Yes | Purchase (30 coins) | ||
77 | Flame Cyan | 32x32 | Yes | Purchase (30 coins) | ||
78 | Flame Green | 32x32 | Yes | Purchase (30 coins) | ||
79 | Flame Red | 32x32 | Yes | Purchase (30 coins) | ||
80 | Flame | 32x32 | Yes | Purchase (30 coins) | ||
81 (T5) | Flaming Pointer | 32x32 | Yes | |||
82 (T5) | Gandalf | 32x32 | Yes | Purchase (21 coins) | ||
83 | Glow | 32x32 | Yes | Likely received for 50000 pixels painted | ||
84 | Gun 4 | 32x32 | Yes | Purchase (25 coins) | ||
85 | Rainbow Sketch | 60x60 | Yes | Paint 20000 pixels Each can holds 1000 pixels so you need to empty your can 20 times | Group 1 | |
86 (T5) | Dark Blue | 32x32 | Yes | Purchase (40 coins) | ||
87 | Pink Heart | 32x32 | No | Purchase (25 coins) | ||
88 | Spray Can | 32x32 | No | Paint 10000 pixels. Each can holds 1000 pixels so you need to empty your can 10 times | Group 1 | |
89 | Trancebear | 32x32 | Yes | Purchase (67 coins) | ||
90 | Blue Sparks | 32x32 | Yes | Paint 33000 pixels. Each can holds 1000 pixels so you need to empty your can 33 times | Group 1 | |
91 | Pink Sword | 32x32 | Yes | Purchase (25 coins) | ||
92 (T6) | Knife | 32x32 | No | |||
93 (T6) | ??? | 32x32 | Yes | |||
94 | TAMMY | 32x32 | Yes | Purchase (67 coins) | ||
95 (T6) | Water | 32x32 | Yes | |||
96 (T6) | Dust | 32x32 | Yes | |||
97 | Basketball | 32x32 | Yes | Purchase (67 coins) | ||
98 | Paw | 32x32 | Yes | Purchase (40 coins) | ||
99 | Dancing Banana | 32x32 | Yes | Purchase (67 coins) | ||
100 | Ghost Hand | 32x32 | Yes | |||
101 | Ghost Cursor | 32x32 | Yes | |||
102 | Frog | 32x32 | Yes | Purchase (67 coins) | ||
103 | Blue Clouds | 32x32 | Yes | |||
104 | Cursed Diamond Sword | 32x32 | Yes | |||
105 | Purple Electric | 32x32 | Yes | Purchase (150 coins) | ||
106 | Fractal Starburst | 32x32 | Yes | |||
107 | Dark Electric | 32x32 | Yes | Purchase (150 coins) | ||
Added 3/6/26 JST – oh no, there's more and they cost alot… | ||||||
108 (T4) | Cow | 32x32 | No | Group 3 | ||
109 (T5) | Angel | 32x32 | Yes | Purchase (100 coins) | Group 3 | |
110 (T5) | Cigarette | 32x32 | Yes | Purchase (100 coins) | Group 3 | |
111 (T5) | doggi | 32x32 | Yes | Purchase (100 coins) | Group 3 | |
112 (T5) | Humming | 32x32 | Yes | Purchase (100 coins) | Group 3 | |
113 (T6) | GlitterFire | 32x32 | Yes | Purchase (500 coins) | Group 3 | |
114 (T6) | peng | 40x40 | Yes | Purchase (500 coins) | Group 3 | |
115 | GlitterSkull | 32x32 | Yes | Purchase (1000 coins) | Group 3 | |
116 (T7) | RainTrail | 32x32 | Yes | Purchase (1000 coins) | Group 3 | |
117 | SmokingUni | 32x32 | Yes | Purchase (1000 coins) | Group 3 | |
118 (T7) | YY | 32x32 | Yes | Purchase (1000 coins) | Group 3 | |
119 | DOLLA | 32x32 | Yes | “Awarded to the best of the best” Be Top 25 in the leaderboard (?) | Group 3 | |
New cursors June 5th, 2026 | ||||||
120 (T4) | Cabbage | 32x32 | No | Purchase (250 coins) | Group 4 | |
121 (T4) | Dice | 32x32 | No | Purchase (250 coins) | Group 4 | |
122 (T4) | Donut | 32x32 | No | Purchase (250 coins) | Group 4 | |
123 (T4) | Slurp | 32x32 | No | Purchase (250 coins) | Group 4 | |
124 (T4) | Spam | 32x32 | No | Purchase (250 coins) | Group 4 | |
125 (T4) | Yuna | 32x32 | No | Purchase (250 coins) | Group 4 | |
126 (T4) | Twist | 32x32 | Yes | Purchase (250 coins) | Group 4 | |
127 (T4) | Uni | 32x32 | No | Purchase (250 coins) | Group 4 | |
128 (T5) | CLICK | 32x32 | No | Purchase (500 coins) | Group 4 | |
129 (T5) | CUBE | 32x32 | No | Purchase (500 coins) | Group 4 | |
130 (T5) | Kunai | 32x32 | Yes | Purchase (500 coins) | Group 4 | |
131 (T5) | Spider Web | 32x32 | No | Purchase (500 coins) | Group 4 | |
132 (T3) | Banana Gun | 32x32 | Yes | Purchase (5000 coins) | Group 4 | |
133 (T4) | BlueBlaze Sword | 32x32 | Yes | Purchase (5000 coins) | Group 4 | |
134 (T5) | LOVE | 32x32 | Yes | Purchase (5000 coins) | Group 4 | |
135 (T4) | Pink Glitter Weed | 32x32 | Yes | Purchase (5000 coins) | Group 4 | |
136 (T3) | PurpDance | 32x32 | Yes | Purchase (5000 coins) | Group 4 | |
137 (T5) | Soulja | 32x32 | Yes | Purchase (5000 coins) | Group 4 | |
138 (T6) | Zenith | 32x32 | Yes | Group 4 | ||
139 (T6) | BLAST | 32x32 | Yes | Group 4 | ||
140 (T6) | Bunny | 32x32 | Yes | Group 4 | ||
141 (T6) | CHEM | 32x32 | Yes | Group 4 | ||
142 (T6) | Double Cup | 32x32 | No | Group 4 | ||
143 (T6) | Glitch | 32x32 | Yes | Group 4 | ||
144 (T6) | Glitter Arrow | 32x32 | Yes | Group 4 | ||
145 (T6) | God Hand | 32x32 | No | Group 4 | ||
146 (T6) | Heaven | 32x32 | Yes | Group 4 | ||
147 (T6) | LSD | 32x32 | Yes | Group 4 | ||
148 (T6) | Mod Hand | 32x32 | Yes | Group 4 | ||
149 (T6) | H4XX0R | 32x32 | Yes | Group 4 | ||
150 (T6) | Scorp | 32x32 | No | Group 4 | ||
151 (T6) | Y2K Angel | 32x32 | Yes | Group 4 | ||
152 (T6) | Cookie Clicker | 32x32 | No | Enter the “Cursor Inventory” and press the cookie 200 times. Refresh the page if it doesn’t show up. | Group 4 | |
🟦 Socials Purge
All of filian’s socials have been purged of content. Tabs related to this will go here.
🟦 Filian's Steam Account
Disclaimer: There is no known clue pointing to Filian’s steam account and it is possibly not part of the ARG or a huge red herring. Double Disclaimer: Linking any Steam account in the Discord gets you a 24 hour mute.
Random_Jawa/@random_jawa found profile history at https://steamladder.com/profile/76561198183932810/rep/, Pizzabash noted the account had been privated Dec 12, 2025, see Figure A1, and the profile name had changed to a single period “.” on Apr 2, 2026 at 9:03PM (Todo get timezone), see Figure A2, and the PFP had been changed on Apr 4, 2026 (Todo get timezone), see Figure A3. ARP_Kioshima/@arp_kioshima found a history of the profile including profile names on https://steamhistory.net/id/76561198183932810, see Figure A4. ARP_Kioshima and TH3EEuphoricFan1161 both verified that a manual indexing of the steamhistory.net site is required for a profile status to be captured. Later Liquid/@liquidwater2 brought up concerns about old profile names missing when looking at the steam account in steam. Using a test account set up by TH3EEuphoricFan1161, Liquid and TH3EEuphoricFan1161 were able to recreate the state of Filian’s official steam account’s profile names, see Figure A5. Note, this took a little finagling to recreate the state, see the filian-arg-read-the-pins channel on Apr 16/Apr 17 to see more details.
Figure A1: Screenshot from Pizzabash showing profile visibility history.
Figure A2: Screenshot from Pizzabash showing profile name change date and time (todo get timezone)
Figure A3: Screenshot from Pizzabash showing PFP change date.
Figure A4: Screenshot from ARP_Kioshima showing the official Filian steam account history https://steamhistory.net/id/76561198183932810 |
Figure A5: (Top Left) Official Filian Steam profile. (Top Right) Official Filian Steam profile showing previously played as profile names. (Bottom Left) TH3EEuphoricFan1161’s test steam profile showing normal profile name change history. (Bottom Right) TH3EEuphoricFan1161’s test steam profile showing a recreation of the Official Filian Steam profile’s profile name state.
🟦 Profile Picture Analysis
On Apr 7, 2026 at 10:19PM UST (Apr 8 in some timezones) “TheSlow / mc: The_Kaktuzz”/@.theslow on the Snackers HQ discord noted the official Filian steam profile had been updated with a profile picture (PFP) that looked black but contained “seriously weird artifacts going on”. After a couple days, It was initially discounted but brought back up on Apr 15 at 6:47PM UST (Apr 16 in some timezones) by Hiro_Chronos ඞ/@based_hiro. Shortly after Metakimi/@metakimi posted the steam PFP with its intensity increased. Later WolfVerzockt/@wolfverzockt posted another version of the PFP with a different intensity used in the pinned messages by Pizzabash/@pizzabash. The original PFP, Metakimi’s, and WolfVerzockt’s images are shown below in Figure 1.
Figure 1: (Left) Original Filian’s Steam PFP. (Middle) Metakimi’s intensity increased version. (Right) WolfVerzockt’s intensity increased version.
The steam profile in question has been proven to be Filian’s official steam account. See Appendix I for a more detailed history of confirmation.
Figure B1: 8x8 grid of saturated and brightened Steam PFP.
Figure B2: Every color channel and arrangement tried.
Red | Green | Blue |
Red/Green | Red/Blue | Green/Blue |
Red/Green/Blue | OG Grid in grayscale | Grayscale + Brightness |
Figure B3: Using aperisolve.com.
Figure B4: (Left) Steam PFP modified pixels. (Right) Filian’s eye compressed in the same way as the Steam PFP, made as a pfp on steam, took out and highlighted all modified pixels.
Figure B5: All modified versions (layers) of the Filian eye compressed and made as a pfp on steam.
Figure B6: Color change of Steam PFP to see colorshifts easier.
Figure B7:
BACKGROUND COLORS - White
FOREGROUND COLORS (1) - Red, Green, Blue, Dark gray, Light gray
FOREGROUND COLORS (2) - Yellow (red + green), Pink (red + blue), Aqua (green + blue), Dark green (green + dark gray), Light red (red + light gray), Dark red (red + dark gray)
FOREGROUND COLORS (3) - Orange (?? + yellow), Purple (?? + yellow), Light yellow (light gray + yellow), Brownish-blue (blue + yellow), Brown (dark red + yellow).
Figure B8: The original 8x8 grid image with a different solve attempt.
🟫 RGB Histogram Analysis
By taking the red (r), green (g), and blue (b) color channel values for each pixel in the image we can produce a histogram of the color values for each color channel. If a potential source image was simply darkened to produce the official Filian Steam account PFP, when linearly scaled to the channel ranges found in the PFP, the histograms should match to a good degree. Below are some commonly suggested source images compared against the PFP. In all, the left-most column contains the rgb histograms for the PFP, the right-most column contains the rgb histograms for the suggested source image, and the middle column contains the rgb histograms for a linear darkening of the image such that the maximum value before darkening matches the maximum value in the PFP after darkening. Analyses done by TH3EEuphoricFan1161.
Figure A6: (Top) Proposed source image: @isitreallyover ARG youtube channel PFP The Filian ARG Document. (Bottom) Comparison of the with the official Filian Steam PFP. No good agreement is found between the histograms → this is not a good candidate for the source image. |
Figure A7: (Top) Proposed source image: OLD official Filian Steam PFP. (Bottom) Comparison of the with the official Filian Steam PFP. No good agreement is found between the histograms → this is not a good candidate for the source image. |
Figure A8: (Top) Proposed source image: Filian YT channel PFP cropped to the right eye region. (Bottom) Comparison of the with the official Filian Steam PFP. No good agreement is found between the histograms → this is not a good candidate for the source image. |
🟧 Theories
Theories:
So what if Filian is an AI?
Do any of the glitches make a MIDI arrangement?
CHANNELS ON STREAM COMMENTING MAY BE LINKED TO ARG
Check Wind Audio, could be stretched song, common ARG tactic.
This is likely to lead to Filian getting a new model.
It sounded too consistent to be a song ngl
^Incredibly slowed audio always sounds very consistent
^^True, but it could be a low ambient track.
^^^Doesn’t hurt to check
-- wind and water audio from the first stream has been sped up and doesn't sound like anything. Water at 10x speed sounds more watery, wind has nothing.
There was some weird beeps and boops might be morse code or something
^ speed up 50x the last audio (low ambiance (ominous)). Does it sounds like the cicadas?
Song Lyrics?:
Time is up
Is it really over
one day..
in another life?
Maybe?
'''End of Minecraft Marathon, distorted outro video played. v=yKYgQrogPEI hidden in distortion.
https://www.youtube.com/watch?v=yKYgQrogPEI leads to unlisted video on a different channel, @isitreallyover or “Time is up”
Channel description is: vul khf... pu huvaoly spml? thfil?
Using Stick7 (an increasingly common catchphrase during the marathon) as a clue we can shift this 7 digits as a Caesar Cipher to get: one day... in another life? maybe?
Channel PFP has hidden HEX code in what appears to be an eye: 69 20 74 68 6f 75 67 68 74 20 69 20 68 61 64 20 6d 6f 72 65 20 74 69 6d 65
Which decodes to: i thought i had more time
The channel started livestreaming with various low volume sounds. Stream description eventually changed to: 01101001 01110100 01110011 Which in binary is: its
TITLE CHANGE: 325113155552413553
With a polybius cypher it wields: MVCEZWQPX
When used in a vigenere-cipher with the key “cryptii” it yields “Keep Going” (found by ryab4ique on discord)
Looking back through everything I'm going to say there's a reasonable chance this isn't a full ARG. We haven't really had clues lead to other clues and most confirmed solutions were just rudimentary codes. I think the stream is supposed to be more of a teaser than any insane hidden codes considering nothing has come of any audio analysis. Even the dial tone might have just been a default sound used for this. I think all the sound effects together are just some sort of teaser to a new model that will make a lot more sense on debut
“The Yuuki Theory” - liquidwater2 on Discord (Initially posted 12/11/25)
—----------------------------------------------------------------------------------------------------------------------------------
summary:
This theory posits that Filian is like Yuuki from SAO. She's actually on her death bed and has been piloting a virtual person. Time is up means she finally croaked and now her consciousness has been transported to the digital world. The reason she hasn’t returned yet is that there was a system crash which corrupted her, which is why she isn’t around right now. We are currently searching for a backup to restore her.
Edit: I think it might be more likely that she is trying to escape something (could still be death) by transferring her consciousness, or at the very least someone doesn’t want her to transfer, which explains the knocking, and why she’s running in the ARG intro clip, and seems to be transferring willingly. Someone also mentioned in another theory that backslash is the escape character in a lot of programming languages, and backslash has shown up a lot in the arg so far.
She has recently removed the “vtuber in training” from her socials, and has been insisting on streams that she is not a vtuber, but instead an anime girl.
The channel description says “In another life”, implying that a current life is ending.
Potential narrative explanation for the 114 backslashes and the underscores:
11 represents Filian because of her jersey. 4 represents death, as it does in japan and other asian cultures. This would mean Filian has died. In the first livestream, one of the identifiable noises was a heart rate monitor with a dangerously slow heartrate. If we assume the backslashes represent Filian’s death, the 28 underscores in the third stream would be a heart rate monitor flatline. The silence is a moment of silence for the dearly departed, and the following music “clair de lune” is common funeral music. The music in the fourth stream is titled after a tragedy where the main heroine killed herself. —-To be expanded upon.
In the fourth ARG livestream we get a repeated code with the pattern 6:22:22. At exactly that second in the minecraft stream filian dies. The odds of this happening are 0.005%, or 1 in 20,352 making it unlikely to be a coincidence.
//Knocking could potentially be “knocking on heaven’s/death’s door”. This is still pretty shaky.//
//The 3rd livestream, the one with silence and funeral music with the supposed flatline, was public for almost exactly 3 days, which could symbolize a period of death before a resurrection//
//The third video has the character limit of forward slashes. Some people have pointed out that the backslashes could be a descent, the underscores the flatline, and the forward slashes a meteoric rise. Unclear if this is the intent//
Supporting evidence:
Video example of this from 2 months ago:
https://youtu.be/Z5CC1MmBJVE?t=2954
The only reactions with substance from the minecraft stream:
Part 5: In response to a glitch, she says “I’m not long for this world”.
4:26:45 | Glitch - “abelebelebele, I think this has gotta be the run, I am not long for this world” |
Part 7: In response to a glitch she says “It’s not lookin’ good”
7:27:45 | Glitch - “oh frick I’m outta time guys, I’m really glad we finished it in 4 days cause it’s not looking good” |
Yuuki Theory Expansion #1 - 2/22/26 (might be a bit of schizo)
This is a quick update to the theory which incorporates new information that has been gained in the past few months. One major question that the previous segment fails to answer is “If she transferred to digital why is she still gone? Why isn’t she back yet?”. She never woke up from the data transfer. The process was incomplete. This is because the system/network transferring her data crashed mid-way through. We know that the windows 95 environment we have access to went through a crash. We have several corrupted files in the recycling bin, a few of which are confirmed to be created when a crash takes place. Filian’s data was also corrupted during this crash. The people facilitating this transfer needed to restore the missing data, so they started extracting her personality and memories from her youtube channel, which is why hundreds of videos and all posts are gone, and why it is getting darker/staticy. The question “any updates on the girl” asked by J is referring to whether she has woken up or been restored by the extracted data. Us snackers on the other hand have an alternate route to a solution. An unknown entity has offered us aid. “If you want your fallen a sacrifice must be made”. While at the time this seemed to be an unplanned helping hand offered to us by the gm due to our lack of minecraft ability, I believe there is more to it. Filian is counted among our fallen by this entity. Unfortunately, while our character data was easy to find and restore, Filian’s is corrupted. It requires a backup. To fulfil the boon we purchased from it and restore our final fallen, the entity provided us a riddle not for backup coordinates in the sense that the original coordinates were lost, but coordinates TO a backup of Filian. This is why the read me includes the line “Sacrifice.complete=True” as though it was a requirement to acquire the riddle.
(short timeline for reference)
1/22/26 330 videos are removed from the filianislost channel
2/11/26 J and Me Convo drops
2/16/26 303 more videos and all posts are removed from the filianislost channel.
Additional speculation (shaky): The streams at the start of the ARG were the data stream of filian’s consciousness being transferred. The messages in the titles and description were not necessarily to us, but thoughts by filian as the process took place. “Keep going, it’s happening, closer”, however the final of these began to show some cracks as it was displayed backwards. During that stream we got the message “Filian Is Lost”, and as mentioned above the following stream had no message. Only a flat line of underscores, silence, and funeral music.
It could also be that Filian is the one that initiated this transfer against the will of whoever is operating this technology, and the knocking is them trying to get in and stop her.
Narrative open ends:
Why did we need a sacrifice/what’s with all the altar and temple iconography?
What is J and Me’s actual goal? Are they antagonists?
Are the backup and “it” the same thing?
BTW if this pretty much correct but the data transfer was the other way from digital to human I’m still gonna count this as being right XD
—----------------------------------------------------------------------------------------------------------------------------------
January 2025 Japan streams:
2025-01-04: 🍊 CHAT SPENDS $1,000 ON ANYTHING IN JAPAN 🍊 IRL in Japan shopping! Using Hollywood tracking today! Also some AR stuff! [Day 2/4 in Japan]
First time going into some details about Debut.
2025-01-13:
First collab back is with Vedal/Neuro
2025-01-24: 🍎 HARDCORE MINECRAFT 🍎 If I die, my world gets deleted. And so does the stream. Goal's to do stuff besides try to beat game. I want a dog
2025-01-29: 🍉[NEW TECH] If I Stop Pedaling My Router, I Die 🍉 Today I hooked my router up to a bike. If I stop pedaling, it turns off. New graphics ya
2025-05-26: https://www.youtube.com/watch?v=6cu7XqnIlCg
2025-09-01: https://youtu.be/tAegXiWfaB4?si=ft5XqqYeBZTCyD7A&t=774
Strange yapping about
Could this mean she signed a significant contract like buying a property or founding a company?
2025-09-01: https://youtu.be/5eUQvDYkfDw?si=a27DTHoSZxEOyrgR&t=12772
2025-09-13: 🍋 YOU LAUGH, YOU LOSE 🍋 IF I LAUGH, I RAW DOG A BOWL OF MUSTARD. I don't like mustard... Good thing I don't laugh easily! Shoutout to Kend
Massive concentration of ARG related happenings:
2025-10-06: https://x.com/filianIsLost/status/1975306378681213346
We can assume something caused the HcMinecraft stream (which likely was planned as start to the ARG and debut) to be pushed back for 2 months. The following 2 months of stream seem slightly derivative, like filler content.
2025-12-02: VSoccer
Distance from mention (2025-09-01) : 92 days.
Distance from planned HcMc Date (2025-10-10): 53 days
Initially on the Livestream 2 page uncredited:
The 114 ‘/’ Theory:
The song is “Aria Variata, BVW. 989 - Variation No. 1” The only other thing that related the number 114 to this song is “Minuet in G major, BWV Anh. 114”. The song was originally written by Christian Petzold, but was popularized by S.Bach. This song is in G Major which might be a hint later on idk but it’s gotta be related to the fact that the title of the livestream was Minor A.
I don’t yet really know how this relates but the words Minuet or G major or Petzold(OG creator) or Anna Magdalena(The piece is part of the notebook that Bach wrote for his wife) might be important.
"Why Backslashes? The 'Escape' Theory" The use of 114 backslashes (\) is not random. In programming, \ is the Escape Character, suggesting the entity is trying to 'break out' of the 114 (The Bach Minuet). Also \ is used for Windows Local Paths, while / is used for Minecraft and URLs. This confirms the entity or Filian is trapped in the System OS, not the game, supported by the 'Windows Ding' sound heard at 49:48." (This is just my theory tho, might be completely wrong)
A variation on this is that 114 is decimal in ASCII for ‘r’ and ‘\r’ is a Carriage Return special character. This is used in Windows to denote a newline in the sequence ‘\r\n’ (‘Carriage Return’ followed by ‘Line Feed’). Reference: https://en.wikipedia.org/wiki/Newline
Music Note Theory:
If you translate 'Find it' into german it becomes 'Finde es'. This has a very interesting double meaning as the word es for it also is the name for the note e-flat in german.
@random_jawa Incidentally the note e-flat is right to the left of the keys E and F, 2 white keys/2 ivory towns on a piano:
@ratinaselfmademaze here's the frequency of it/e-flat. Maybe in some kind of recording there is something interesting at that frequency.
I also do not know where else we shout find e-flat but it also appears a few times in the right hand of Claire-de Lune by Debussy so maybe something happens at those times in the recording, or we just have to count how often e-flat appears or smth. Similar.
And lastly to give fuel to the conspiracy schitzos among us: The mods who totally do not know the solution and do not give tips * wink * * wink * posted in the arg channel that e-flat stands for filian so by finding it/e-flat we can find filian:
🟧 "find it" Theories
wSillyProotSoda and Veter992:
Thinking that “it” could be a item in suspicious sand or gravel, going to recheck trial chambers for them.
Makedonski:
I have a schizo theory. I rechecked the recording with minecraft steps, and with a big gain, it can be seen that this is a poorly trimmed cycle of 6 slow, 11 normal and 10 fast steps. I thought it was logical to convert it to time (the slower the step, the longer the time) like 6:11:10. I decided to check the VODS, and found that in the first stream, perfect at this timing, the neuro disappears from the server https://youtu.be/RksF7wbZgfA?t=22270
https://www.twitch.tv/videos/2636167728?t=6h13m3s
Summary:"find it" = "find neuro"
How do we find neuro? solve the neuro arg
aaLinnea:
Time.
Time has been an overarching theme since the beginning, finding timestamps, finding dates, filian “running out of time” “time is up” “thought I had more time”. What have we lost? What is lost? Time. What are we searching for? … Maybe a timestamp. Maybe the timestamp of a debut.
Apton_ika:
In older clips she played the account ‘FilianIsLost’ (as in the ARG).
https://minecraftuuid.com/player/944da1d5-718f-4527-b37a-b454d550b217
In the actual HC Minecraft run she played on ‘FilianIsCringe’
https://minecraftuuid.com/player/70916122-8e54-4f4b-a800-6042935e650c
Fog_runner:
Possible relevant info from “diamond city lights” dance short, that had the link to arg discord server.
Based on the lyrics I have a couple of ideas.
1. «Data behind my heart's encryption key (Can you see?)
Listen and you might hear a melody»
Either the number of heartbeats (49) will be relevant soon, or it’s talking about the “heart” shape we should be looking for. The map people posted in “Minecraft server” has at least 2 candidates.
2. «When gray skies take the colors away (Far away) I'll paint a rainbow lasting for days and days», «Rainbow-colored skies invite us to take a ride», «A shimmering light that will guide you through the night». «In the night there's no need to hide».
Maybe clues are set to appear at night cycle or they are more obvious at night. Gray sky areas could be more relevant to the arg than others. Gray skies are common in winter, and the resurrection altar was found in the winter biome surrounded by flames that are easier to see at night. Maybe this clue is now exhausted.
3. «Let's go to the place», «Diamond city lights illuminate what's inside», «Decoding these emotions, you're the ocean I wanna dive into tonight».
This combined with point 1 suggests that there might be something beneath a heart-shaped body of water. Or, if heart shape is irrelevant, beneath any big body of water that could pass as an ocean by Filian’s logic.
If I unleash my schizo interpreter, then there might even be music of sorts somewhere deep in the water to signify a correct place (a hidden jukebox maybe?). But based on the clues we had previously I wouldn’t rely on this.
Cyrus:
If you use chappe code on the server separated into digits with the username _._.__.___.____.___.__._._ then you get:
décès graduer ? craindre que folio
which translates to:
death graduation ? fear that folio
This is likely in reference to the file on blackfile index meaning “it” will likely be a password to CITF_MEMO.TIF
It is worth noting that 9999 in the middle of the server seed does not translate to chappe code, hence the question mark.
Believed to be a one-time resurrection that pulsed for a time, it has subsequently been destroyed by the community in pursuit of knowledge.
It was initially thought that the End Shrine was the “it” that we were looking for. The shrine has since been heavily damaged. However, the server description changed and then changed again to emphasize “find. it.” indicating otherwise.
Similar to other structures, this was thought to be “it” but that seems unlikely at this point.
🟦 Misc Info
🟨 Unused Numbers
7.83 | difference in l/r frequencies, "schumann resonance" frequency |
111#D345A | or 111#1D345A or other possible variations |
49 | heart beats |
4,11,10,6,11,10,6,11,10,1 | minecraft steps: slow, normal, fast. 80 total. |
114 | backslashes in description |
989 | bach musical work number |
28 | underscores in title |
75 | debussy musical work number |
3 | third movement of the musical work |
28 | underscores in title |
4999 | slashes in description |
50 | musical work number |
3 | third part of the musical work |
7 | 7 | 7 | (or 4 1 2 | 4 1 2 | 4 1 2) typewriter pattern (rare exception to 6 | 2 2 | 2 2) |
24 | slashes in discord server name |
1.1.2.3.4.3.2.1.1 | pattern of underscores in discord creator’s username |
711026799995424918 | Minecraft server seed |
🟦 Hardcore Video Links
MC Marathon runs timestamps (made by Makedonski):
1 https://youtu.be/RksF7wbZgfA?t=1439
2 https://youtu.be/RksF7wbZgfA?t=1647
3 https://youtu.be/RksF7wbZgfA?t=3810
4 https://youtu.be/RksF7wbZgfA?t=3902
5 https://youtu.be/RksF7wbZgfA?t=4370
6 https://youtu.be/RksF7wbZgfA?t=4841
7 https://youtu.be/RksF7wbZgfA?t=5602 frozen similar to seed?
8 https://youtu.be/RksF7wbZgfA?t=6760
9 https://youtu.be/RksF7wbZgfA?t=7444 frozen
10 https://youtu.be/RksF7wbZgfA?t=8214
11 https://youtu.be/RksF7wbZgfA?t=11286
12 https://youtu.be/RksF7wbZgfA?t=11962
13 https://youtu.be/RksF7wbZgfA?t=13301
14 https://youtu.be/RksF7wbZgfA?t=14031
15 https://youtu.be/RksF7wbZgfA?t=20212
16 https://youtu.be/RksF7wbZgfA?t=21736
17 https://youtu.be/RksF7wbZgfA?t=25754
18 https://youtu.be/RksF7wbZgfA?t=26444
19 https://youtu.be/RksF7wbZgfA?t=27624 frozen
20 https://youtu.be/RksF7wbZgfA?t=27760 frozen similar to seed?
21 https://youtu.be/RksF7wbZgfA?t=30413
22 https://youtu.be/RksF7wbZgfA?t=31129
23 https://youtu.be/RksF7wbZgfA?t=31601
24 https://youtu.be/RksF7wbZgfA?t=32026
25 https://youtu.be/RksF7wbZgfA?t=35018
26 https://youtu.be/RksF7wbZgfA?t=36088
27 https://youtu.be/RksF7wbZgfA?t=37306
28 https://youtu.be/RksF7wbZgfA?t=38338
29 https://youtu.be/RksF7wbZgfA?t=38660
30 https://youtu.be/RksF7wbZgfA?t=41761
31 https://youtu.be/RksF7wbZgfA?t=42214
32 https://youtu.be/wS_XJcZFRJY?t=230
33 https://youtu.be/wS_XJcZFRJY?t=674
34 https://youtu.be/wS_XJcZFRJY?t=1675
35 https://youtu.be/wS_XJcZFRJY?t=40891 frozen
36 https://youtu.be/wS_XJcZFRJY?t=41974
37 https://youtu.be/hET_j3jTdFc?t=562
38 https://youtu.be/hET_j3jTdFc?t=1498
39 https://youtu.be/hET_j3jTdFc?t=5801
40 https://youtu.be/hET_j3jTdFc?t=6017
41 https://youtu.be/hET_j3jTdFc?t=8401 frozen similar to seed ?
42 https://youtu.be/hET_j3jTdFc?t=10459
43 https://youtu.be/hET_j3jTdFc?t=11633 frozen similar to seed ?
44 https://youtu.be/hET_j3jTdFc?t=12631
45 https://youtu.be/hET_j3jTdFc?t=12805
46 https://youtu.be/hET_j3jTdFc?t=13170
47 https://youtu.be/hET_j3jTdFc?t=13551
48 https://youtu.be/hET_j3jTdFc?t=14410
49 https://youtu.be/hET_j3jTdFc?t=15804
50 https://youtu.be/hET_j3jTdFc?t=19034
51 https://youtu.be/hET_j3jTdFc?t=19231
52 https://youtu.be/hET_j3jTdFc?t=26888
53 https://youtu.be/hET_j3jTdFc?t=27743
54 https://youtu.be/hET_j3jTdFc?t=28392
55 https://youtu.be/hET_j3jTdFc?t=29150
56 https://youtu.be/hET_j3jTdFc?t=29905
57 https://youtu.be/hET_j3jTdFc?t=30993
58 https://youtu.be/hET_j3jTdFc?t=32448
59 https://youtu.be/hET_j3jTdFc?t=33033
60 https://youtu.be/hET_j3jTdFc?t=33397
61 https://youtu.be/hET_j3jTdFc?t=34310
62 https://youtu.be/hET_j3jTdFc?t=36350
63 https://youtu.be/hET_j3jTdFc?t=37148
64 https://youtu.be/hET_j3jTdFc?t=37515
65 https://youtu.be/TFexmJK5G4E?t=660
66 https://youtu.be/TFexmJK5G4E?t=898
67 https://youtu.be/TFexmJK5G4E?t=14320
68 https://youtu.be/7g9SDJj1iaI?t=3710
69 https://youtu.be/7g9SDJj1iaI?t=4541
70 https://youtu.be/7g9SDJj1iaI?t=5432
71 https://youtu.be/7g9SDJj1iaI?t=5657
72 https://youtu.be/7g9SDJj1iaI?t=6027
73 https://youtu.be/7g9SDJj1iaI?t=11598
74 https://youtu.be/7g9SDJj1iaI?t=12370
75 https://youtu.be/7g9SDJj1iaI?t=13950
76 https://youtu.be/7g9SDJj1iaI?t=18402
77 https://youtu.be/7g9SDJj1iaI?t=19555
78 https://youtu.be/7g9SDJj1iaI?t=19947
79 https://youtu.be/7g9SDJj1iaI?t=20325
80 https://youtu.be/7g9SDJj1iaI?t=22413
81 https://youtu.be/7g9SDJj1iaI?t=35125
82 https://youtu.be/7g9SDJj1iaI?t=35980
83 https://youtu.be/sgY3n8GGH6I?t=499
84 https://youtu.be/sgY3n8GGH6I?t=1132 frozen
85 https://youtu.be/sgY3n8GGH6I?t=38284
86 https://youtu.be/sgY3n8GGH6I?t=41607
87 https://youtu.be/sgY3n8GGH6I?t=41959
MC Practice runs(made by Makedonski)
1 https://youtu.be/iHLR7e35gQs?t=1193
2 https://youtu.be/xoB4Uf9Gc9I?t=398 died under snow(t=12370)
3 https://youtu.be/Qa_d15dlw24?t=473
4 https://youtu.be/F7IXGGYE444?t=214
5 https://youtu.be/PQeXVUOBjCA?t=293
6 https://youtu.be/PQeXVUOBjCA?t=4680 died to boat bug (t=6652)
7 https://youtu.be/tOpwxaE-3mA?t=314
8 https://youtu.be/9WhBoPAbCXU?t=288 frozen similar to seed?
🟥 Related Memes
This tab is dedicated to memes and jokes that people made in relation to solving this ARG
Frequently recurring meme when chatters return to channel after long break and ask if there’s anything new:
After several days of Claire de Lune:
When Filian joined the ARG Channel:media.tenor.com/Ssfj2DOGJhAAAAPo/filian-vtuber.mp4 (idk how to insert a GIF or a video here)
Regarding user who kept replying to posts more than day old:
Apparently this is what actually happened on 12/19/2025
🟫 Schizo Nonsense
🟫 Probably-nothing????
🟫 vsoccer
The Vsoccer website, as pointed out by ExodusCore in the youtube livestream chat of the third livestream contains potential references to elements of the arg;
https://vsoccershow.com/
(screenshot credits: ARP_Kioshima on discord)
Makedonski:
Checked all site images
Found that:
this means that most likely the number 11 was added during postprocessing, and such a T-shirt does not actually exist.
I have a code that assigns a random color to each color, if the colors are the same, then they had the same color before processing. In real photos, the pixels are of different colors due to noise.
What interesting: it only appears on Filian's green T-shirt, and note that only the number stands out. That is, they had a green Filian T-shirt, but without a number, and a purple Filian T-shirt, but with a number. It's at least weird.
Also, there are only 2 images like model_x_fil:
https://vsoccershow.com/cdn/shop/files/model_1_fil.png
https://vsoccershow.com/cdn/shop/files/model_7_fil.png
Stick Seven ???
Found on 22.12.2025 11:41 GMT:
https://vsoccershow.com/password
(do not the page appeared because the shop closed…)
Note : Not confirmed if this is arg related or not. Probably not.
Note² : You get locked 10mn after too many failed attempts
List of tried Passwords:
2632 | Suggested by @Willis on discord |
FilianIsLost filianislost | |
donttrusti | |
11D345A11, 1D345A1, D345A | |
4999, 711026799995424918 | |
709161228e544f4ba8006042935e650c password | |
filian | |
admin | |
Admin | |
rachel | |
🟫 vids/clips
wwwwwwwThe following is a list of clips or videos that are or may be relevant in the arg.
Minecraft hardcore challenge marathon ending (posted 9th Dec 2025);
Filian Dropped a Teaser at the End of Her Marathon Stream...
2 month old clip(from 13th Sept 2025 stream) of Filian talking about up coming debut;
Filian Gives An Update About Her Debut
Full vod;
Filian's You Laugh. You Lose - Mustard Bowl (9/13/2025)
(also note this was the first HC minecraft practice and first look at wplace)
6 month old clip(from 26th May 2025 stream) of Filian talking about face reveal/up coming model debut;
The New Filian Is Going To Debut Soon...
Full vod;
Filian, Cotton and Mega Compete at Geoguessr (5/26/2025)
First Mention of Debut Planning in January 2025:
🍊 CHAT SPENDS $1,000 ON ANYTHING IN JAPAN 🍊 IRL in Japan shopping! [1/4/25]
https://youtu.be/Bw--JpoT5n4?si=aGXBsO9iHlfkoI9z&t=6983
Sept 1st, Filian says she will show us her closet when she moves to new place (suggests possibly dynamic mapping of rooms/objects or full apartment mapping in so much detail as to include her closet)
https://rumble.com/embed/v6w6wlk/?pub=4&start=1720
Transcripts:
May 26th clip:
when's the face reveal? *big sigh* chat you guys wanna see my real face?!
i'll talk about it. guys, here's some information. as of this morning, based on what i have seen of the future filian i'll say *ahem* i am now 100% confident that debut is happening. it is going really well. it looks amazing.
a little behind the scenes, chat. a lot of you say, "filian why have you not debuted after three and a half years of streaming?"
the answer is simple, from the beginning i've always seen myself as sort of like an anime girl. and that's kind of like what i want to be in the future, right?
so with that being said i tried to debut twice! twice in the past but both times i was not satisfied with it, so i had to put it down
anything less than the best for you guys is unacceptable
so, chat, i can say though after all this time, i really think we're gonna be getting the best this time
it took me three years of meeting people learning and a lot of other stuff but i think for the future you guys will be very very excited
now chat, i will be saying this, "filian, *ahem* what are you doing once the model's ready?"
and the answer is, a lot. there is so much more going into it than just my physical appearance
there's going to be a lot more happening
so you know chat, technically we're still in early stages but i'm really excited for it. i know it's happening, and i also think you guys will be um very impressed, i guess
we'll see, chat, we'll see. anyway snackers thats all i gotta say. no dates, no dates
but as of this morning, from what i've seen, i'm now 100% certain this will be what you guys will see
Sept 13th clip:
so guys, i have a quick thing to say, a lot of you guys have been like, "filian debut when?"
ok i i'll say this, chat, it's probably not as far away as you think
but i will also say this, chat, it's not gonna be a normal debut
it's gonna be probably the first time in like fore- definitely the first time but maybe the last time that anyone's ever done something like this
so it'll be a little bit different, chat, instead of just debuting, i've been preparing to show you guys my- my world
a lot of you ask like, "filian, can i ask you questions like you know uhhh like where do you live and stuff?" you know i'm saying?
and you know i usually just say a fruit snack factory, cardboard box, homeless, blah blah blah
but the truth is, chat, i live in an apartment, and i am going to be moving to a slightly bigger one soon
and then i'll be showing it for you guys. i'll be showing you guys the world. so yeah
i've been working on it nonstop in the background and its been very very intensive so i hope you guys will be excited for it
i am working beyond hard for it. its the hardest i've ever worked for anything in my life
so like doing all the streaming stuff on top of it like a lot of its like you know not that crazy
but yeah, i've been working really really hard on something you know special
we've been waiting four/for years for it so yeah i'll show you guys my world i guess
so yeah that's all that's all i wanted to say uhm no dates no dates no no no no no no no no dates
but when i'm done preparing everything, i'll show you guys.
ok, chat? alright? cool? yes or no? is that ok chat? fighting!
Dec 8th clip:
i didn't think it would be this soo-
IT'S HAPPENING!
it's time for me to go for a- for a little while
i (unintelligible)! (maybe "i'm running out of time!")
i really gotta go!
i didn't think you'd ever see me like this
you'll be able to see the world, my world
🟥 Logged Message (Will)
From 1st Livestream
This section has been debunked. I will not be deleting it, since I assume a good amount of work went into it, but read with several grains of salt:
From this “@Will-u6m-d8i” guy [https://www.youtube.com/@Will-u6m-d8i]
Dunno why, saw a few people note these messages so just keeping them here in case of anything idk
[dababe31 on Discord]
[engirenouf on Discord]
[secretly_doggo on Discord]
[JCML hi yes that’s me .w.]
🟫 Time is up schizo
Initial video:
Theory - Throughout the video the black screen gradually turns to purple. This was likely the main clue in the video. It was saying take the black photo (pfp) and raise the contrast to reveal the code. The code was in Filian’s eye which is purple.
Livestream 1:
🟫 Binaural Sinewaves
Livestream 1
4:52-10:20 | Binaural sinewave left=40hz right=47.83hz Rapid popping at start More popping at 5:45 Extra noises at 7:22-7:51 Popping at 8:31-10:20 |
Binaural sinewaves also known as binaural beats are a pseudo scientific concept associated with areas like spiritual healing and similar alternative health topics. Within the stream were: left=40hz right=47.83hz this is considered a 7hz binaural beat user Stardrop on discord has noted this is a frequency associated with “healing” - that said the 7hz is on the Theta spectrum of brainwaves (4-8hz) and is related to deep relaxation, a meditative like state and creativity. It is also a prominent pattern seen in REM sleep and has been associated with rapid eye movement in sleep.
Further research and analysis on this section has been performed by several users on discord including audio engineers and found no meaningful leads have been found
Additional info researched and provided by Finite Emporium on discord
🟫 DTMF Tones Analysis
Livestream 1
16:56-17:40 | phone dial (DTMF tones) (NEW) Seem to match tones at the start of VHS tapes (see page 5 of this tab for new information) |
Credit to Mateon1st on discord for the original faint tone analysis
(1) is a 1 that lowers pitch halfway through
Full Message High Amplitude:
(1) 1 1#(1)D345 A 1#(1)D345 A (1) 1 1#D345 A
Full Message Low Amplitude:
5678 # 2345678 #
Full Message Combined:
(1) 1 56781#(1)D345#A 1#(1)D345 A (1) 12345678 1#D345 # A
Plain text:
High Amplitude: 111#1D345A1#1D345A111#D345A
Low Amplitude: 5678#2345678#
Combined: 1156781#1D345#A1#1D345A1123456781#D345#A
Left Channel DTMF FFT Analysis (GeoScape on Discord):
First Part Faint Frequency:
If combined we get: 1 1 5678 1#D345 # A 11#1D345 A
Last Part Faint Frequency (Middle part has none):
1 1 2345678 1#1D345 # A
^Toying with the idea of using the difference of the 2 parts of the string as either a stand alone string or a Shift Cypher Key
Authored, FFTScan method, and new found way to decode DTMF → 4bits by TH3EEuphoricFan1161/@TH3EEuphoricFan1161 in the Snackers HQ discord. Special thanks to Antuber/@antuber_46672 for first suggesting the VHS DTMF tones.
Note this is a WIP as I will be adding more, for the moment this should give a decent baseline amount of info on the tone structuring.
General DTMF (Dual-Tone Multi-Frequency) information can be found here https://en.wikipedia.org/wiki/DTMF_signaling.
Below are some common methods used to analyse waveforms. I also include a detailed description of the Fast Fourier Transform scanning method I use as I use it heavily when analysing DTMF sequences.
The Fast Fourier Transform - https://en.wikipedia.org/wiki/Fast_Fourier_transform
Spectrogram - https://en.wikipedia.org/wiki/Spectrogram
zoomFFT - A very fast method to extract only a subset of the full FFT information https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.ZoomFFT.html
(Much less common) Cepstrum - https://en.wikipedia.org/wiki/Cepstrum
This method of scanning a waveform makes it easy to focus on and read DTMF tones and their relative power quickly, including for very faint tones. The scan is produced by the following steps:
The resulting plot for a single frequency window shows an approximate normalized power in that frequency band over a sliding window in time for the audio sample. An example of a singular frequency band and an example of all frequency bands I use is shown in Figure 1.
(Left) Example FFTScan result using the frequency band ±12Hz around the DTMF related frequency 697Hz. (Right) Example FFTScan result using the set of nominal DTMF frequencies with loud DTMF tones labeled by their corresponding keypad character. Audio source https://www.youtube.com/watch?v=h9Mgl307_dI. | |
Figure 1 | |
From my searching, DTMF tones generally fall into two shape categories based on their symmetry around 0 in the waveform. The most common is the symmetric tones, these are seen in phone dialing, radio/transceiver communications, and Audacity’s DTMF generator . The rarer asymmetric tones I have only found in DTMF tones found on VHS tapes. Examples of each can be found in Figure 2.
(Left) Example of a symmetric tone, in this case from a transceiver; audio source: https://www.youtube.com/shorts/chRzHE5B3A8. (Right) Example of an asymmetric tone, in this case from a VHS tape; audio source https://www.youtube.com/watch?v=h9Mgl307_dI. | |
Figure 2 | |
The shape of the asymmetry in DTMF tones from VHS tapes are worth a quick extra note here as they will be relevant to the Livestream 1 DTMF tones. To see the underlying shape of the asymmetry we can remove the two DTMF tone frequency peaks in the Fast Fourier Transform (FFT) and reconstruct the waveform. An example using DTMF Tone “1” can be seen in Figure 3.
(Left) Original asymmetric tone. (Right) Asymmetric tone with DTMF tone “1” frequencies 697Hz and 1209Hz removed. Audio source https://www.youtube.com/watch?v=h9Mgl307_dI. | |
Figure 3 | |
In live broadcasts, 4 consecutive tones are used to cue up commercials on VHS tapes to play, see page 80 of this magazine: https://www.worldradiohistory.com/Archive-All-BC-Engineering/BE/80s/BE-1983-10.pdf
In the production of VHS tapes, DTMF tones were used in some capacity and remain on some VHS tapes after their printing. A Mirror Master VTR would be created from the master tape. A duplicator, like the Sony Sprinter or Otari T-710 machines used at many duplicator facilities such as Technicolor’s Livonia, MI site, then copies the Mirror Master VTR onto a long VHS tape strip making multiple copies. These copies are then separated by the machine that slices and boxes the VHS tapes. The information on this is relatively diffuse and details hard to find, below is a list of links with relevant info (lets add to this as needed).
https://homevideo.wikitide.org/wiki/Technicolor_test_patterns |
https://homevideo.wikitide.org/wiki/Technicolor_DTMF_sequences Note, that while their discussion on DTMF tones is generally accurate, some information on faint tones and latter loud tones are not discussed. |
The DTMF tones on VHS tapes follow well defined ordering and structure. There are usually 15 tones, sometimes there are 16 (this seems to be a very rare case). The first 8 tones count from 1 to 8, they are faint tones with one selected to be a loud tone. In the case there are 15 tones, the latter 7 tones will all be loud. In the case of 16 tones, all remaining tones are loud except for the second to last which is faint. See Figure 4 for a visual representation of the two tone sets using a FFTScan. Note that due to the window/increment structure of the FFTScan, faint tones after sufficiently powerful tones may be not as easy to distinguish as other faint tones (the 2nd tone in the left example and the 7th tone in the right example of Figure 4).
(Top Left) Example FFTScan result on a VHS DTMF 15 tone set. (Bottom Left) Example FFTScan result on a VHS DTMF 15 tone set zoomed in to see the faint tones. (Top Right) Example FFTScan result on a VHS DTMF 16 tone set. (Bottom Right) Example FFTScan result on a VHS DTMF 16 tone set zoomed in to see the faint tones. Left audio source: https://www.youtube.com/watch?v=h9Mgl307_dI, right audio source: https://www.youtube.com/watch?v=mBAaujIxqik. | |
Figure 4 | |
In all VHS examples I have seen thus far:
- The first 8 DTMF tones always map to the keypad counting up “1” to “8”
- The 9th and 10th DTMF tones of always map to “1” and “#” respectively.
- The 11th tone in >90% of cases maps to “D”, in the remaining cases it maps to “1”
- The 12th tone seems to be free to tones “1” through “9” and “D” unless the 11th tone maps to “1” in which case this tone maps to “D”
- The 13th tone seems to be free to tones “1” through “9” and “D”
- The 14th tone seems to be be mostly from tones “1” through “5” but may be free to tones “1” through “9” and “D”
- If there are 15 tones in the set, the 15th tone seems to be mostly from tones “*”, “0”, “#”, “A”, “B”, and “C”. If there are 16 tones, I have only seen “#” (but the statistics for these examples are very low)
- If there are 16 tones in the set, I have only seen “A” (but the statistics for these examples are very low)
A breakdown of the tones present in each location is shown in Figure 5 as a set of histograms. (Note, this pulls from examples on youtube with decent reliability, I still need to add some and would like to make another plot with only the most reliable sources.)
Distribution of 139 VHS DTMF tones in each position in the structure. Note the rarer 16 tone structure is not yet entered into the dataset I am building. | |
Figure 5 |
VHS DTMFs are also found on the left audio track and very rarely the right (they are solely in the left channel in all the examples I have seen). An interesting quirk of VHS DTMFs is the highest frequency at nominally 1633Hz peaks higher at ~1644Hz where other peaks are much closer to the nominal value. This behaviour is not found in other sources. See Figure 6 for an example made by fitting the FFT distribution with Gaussians. The periodic nature of the FFT in Figure 6 is explained well by prominent quefrencies in the cepstrum (see Figure 7) and seems to be a common element of VHS DTMFs (perhaps more generally too?).
(Top) The left waveform’s FFT fitted for all 8 DTMF related frequencies. White lines denote the nominal peak location. (Bottom) Zoom in on the highest frequency peak showing the shift towards higher frequency than nominal. Audio source: https://www.youtube.com/watch?v=mBAaujIxqik. Fit result table may be found in Figure A1 in Appendix I. | |
Figure 6 |
Quefrency peaks in the cepstrum of the left waveform. Audio source: https://www.youtube.com/watch?v=mBAaujIxqik. | |
Figure 7 |
The waveforms for left, right, and left-right difference are shown in Figure 8. The DTMF tones contain two distinct varieties I label as VHS-type Tones and Shifted Tones. VHS-type tones have the features described above in the DTMF Tone Shape in Waveforms section. These VHS-type tones can be organized into 3 blocks of the rare 16 tone structure described above in the DTMF Tone Structures section. The shifted tones begin as the nominal DTMF tone “1”, but around mid-way through the tone shifts down to a non-DTMF tone with frequency peaks at around 655Hz and 1142Hz. These values are found by fitting the FFT distribution with Gaussians with an example result in Figure 9. In FFTScan plots, the shifted tones can be found by a peak with tone “1” followed closely by a peak of the two shifted frequencies. An example of this can be found in the largest amplitude peak in Figure 10 top or middle.
(Top) The left waveform. (Middle) The right waveform. (Bottom) The left-right difference waveform. | |
Figure 8 |
(Top) The left waveform’s FFT fitted for all 8 DTMF related frequencies and 2 shifted frequencies. White lines denote the nominal peak location. (Middle Top) Zoom in on the highest frequency peak showing the shift towards higher frequency than nominal as expected in VHS-type tones. (Middle Bottom) The right waveform’s FFT fitted for all 8 DTMF related frequencies and 2 shifted frequencies. (Bottom) The left-right difference waveform’s FFT fitted for all 8 DTMF related frequencies and 2 shifted frequencies. Fit result tables may be found in Figure A2 in Appendix I. | |
Figure 9 |
(Top) FFTScan of the left waveform. (Middle) FFTScan of the right waveform. (Bottom) FFTScan of the left-right difference waveform. | |
Figure 10 |
In examining the plots above, we can conclude the following:
These three repetitions I will refer to as VHS DTMF Blocks, and I highlight them in the left-right difference in Figure 11 (Top). We can further note that the shifted tone that precedes a VHS DTMF Block remains a consistent distance in time away from the VHS DTMF Block. I will expand these to include the preceding shifted tone and simply refer to these as Full Blocks, highlighted in the left channel in Figure 11 (Bottom).
(Top) Left-right difference FFTScan with VHS DTMF Blocks labeled. (Bottom) Left channel FFTScan with Full Blocks labeled. | |
Figure 11 |
We may also note the faint tones in Figure 12. All faint tones expected from a VHS DTMF set are present in VHS DTMF Block 3. In VHS DTMF Block 1, faint tones “5”, “6”, “7”, and “8” are fully present while “4” is decently reduced and “2” and “3” are not visible above noise if they are present. In VHS DTMF Block 2, no faint tones are present above the level of noise.
(Top) FFTScan of the left channel, zoomed in on the y-axis to see the faint tones. (Bottom 6) Faint tones for each VHS DTMF Block, block number labeled in plots. | |
Figure 12 | |
Also of note is the waveforms of individual DTMF tones which have multiple breaks leading and following the main tone body, see Figure 13. This can be explained by tempo slowing the audio sample in a program such as Audacity (but reproducing it is a bit finicky, feel free to try to get a better replication than I have).
(Top) Zoom in on a VHS-type tone in the left channel. Obvious breaks and a potential break are noted. (Bottom Left) Example VHS-type tone in Audacity no tempo slowing. (Bottom Right) Same example VHS-type tone in Audacity with tempo slowing. (Bottom) Another example produced by stretching a set of tones from a VHS multiple times with a tempo slow set to -20 in Audacity, audio src: https://www.youtube.com/watch?v=mBAaujIxqik. | |
Figure 13 | |
A straightforward choice of decoding DTMFs is using the standard convention used by DTMF decoder chips such as M8870, HT9170B, HT9170D, CM8870, CM8880 etc. The mapping is shown in Table 1. The loud tones in our base VHS DTMF Block would decode to 11C0345D in hex, and including faint tones in parentheses assuming all faint tones are present would decode to 1(2345678)1C0345(C)D. The technical document for the M8870 describes this mapping on page 5: https://people.ece.cornell.edu/land/courses/ece4760/FinalProjects/s2001/pr57/M-8870.PDF
DTMF Tones | Keypad Char | Individual Bits | Binary String | Hex | |||
Q4 | Q3 | Q2 | Q1 | ||||
697Hz & 1209Hz | “1” | 0 | 0 | 0 | 1 | 0001 | 1 |
697Hz & 1336Hz | “2” | 0 | 0 | 1 | 0 | 0010 | 2 |
697Hz & 1477Hz | “3” | 0 | 0 | 1 | 1 | 0011 | 3 |
770Hz & 1209Hz | “4” | 0 | 1 | 0 | 0 | 0100 | 4 |
770Hz & 1336Hz | “5” | 0 | 1 | 0 | 1 | 0101 | 5 |
770Hz & 1477Hz | “6” | 0 | 1 | 1 | 0 | 0110 | 6 |
852Hz & 1209Hz | “7” | 0 | 1 | 1 | 1 | 0111 | 7 |
852Hz & 1336Hz | “8” | 1 | 0 | 0 | 0 | 1000 | 8 |
852Hz & 1477Hz | “9” | 1 | 0 | 0 | 1 | 1001 | 9 |
941Hz & 1336Hz | “0” | 1 | 0 | 1 | 0 | 1010 | A |
941Hz & 1209Hz | “*” | 1 | 0 | 1 | 1 | 1011 | B |
941Hz & 1477Hz | “#” | 1 | 1 | 0 | 0 | 1100 | C |
697Hz & 1633Hz | “A” | 1 | 1 | 0 | 1 | 1101 | D |
770Hz & 1633Hz | “B” | 1 | 1 | 1 | 0 | 1110 | E |
852Hz & 1633Hz | “C” | 1 | 1 | 1 | 1 | 1111 | F |
941Hz & 1633Hz | “D” | 0 | 0 | 0 | 0 | 0000 | 0 |
Table 1 | |||||||
Fit results for the FFT fits shown above.
Fit results for the fits shown in Figure 6. | |
Figure A1 |
Fit results for the fits shown in Figure 9. | |
Figure A2 |
🟫 Heart Rate Monitor
Livestream 1
7:11:22-7:12:52 | Beeping (heart monitor?) 49 beeps audible, a loop of 15 beeps repeated |
Stock audio clip: https://www.youtube.com/watch?v=yEb4kojtVHw The video even has the same volume decrease and increase as the livestream. So this heart rate monitor probably has no encoding info.
[source: glaukopes (not a doctor just an experienced endurance athlete)]
49 beats, but the 49th is very faded
duration 1:27 or 87 seconds
~33 bpm
33 bpm is abnormally low. A "normal" resting heart rate is usually said to be 60-80, but endurance athletes often have resting heart rates in the 50's or even 40's. Even for an endurance athlete, 33 would be extremely low.
🟫 Audio Glitch Analysis
LIVESTREAM 2
Audio glitch aprox 44 m 10 s https://youtube.com/clip/UgkxYYfTmx-tF8y0pHKfq3-RfhoOua9Y-7HP?si=xcoFMII4Ke0Z26Vt
These glitches appear to be natural and are presumed to mean nothing.
🟫 Clair de Lune Analysis
Livestream 3
The stream of Clair De Lune contains audio distortion that sounds like clipping or stopping, it was theorized by some users that this may be some kind of encoding like binary or similar code into the audio to solve. User Mateon1st on discord investigated the distortions for alignment across loops of the music and detected clear inconsistencies between the glitching, unable to produce an alignment that suggests its a looping code intended to be discovered below is an attached screenshot of one of Mateon1st’s attempts proving the encoded audio theory highly implausible:
The main glitches are all exactly 1024 or sometimes 2048 samples long, this is the packet size of mp4a/AAC-LC audio that youtube is using for the livestream. All glitches appear at the same relative offset modulo 1024 from the start of stream (all glitches start/end at sample 42 % 1024), which is decent evidence that this glitching originates in filian's audio/livestream setup, and is likely some kind of corruption or packet loss. (see below)
The audio cuts to near silence on both channels during the glitches, sharply cutting off the audio waves, which causes the audible pops. This shows as a straight, featureless vertical bar in spectrograms
As a follow up and verification of the work done by Mateon1st, Liquid on discord investigated the glitches as well. After extensive work to determine if perhaps there was a longer duration loop of embedded audio across several repeats of Clair De Lune. By cross referencing several longer runs of audio/loops to determine if a pattern or alignment was possible, Liquid concluded that there is no larger pattern that signifies the glitches/audio were placed intentionally. Below are 2 screenshots of Liquid’s work:
—--------------------------------------------------------------------------------------
In case it's of importance the poem in both original French and translated english;
(French)
Clair de lune Votre âme est un paysage choisi Que vont charmant masques et bergamasques Jouant du luth et dansant et quasi Tristes sous leurs déguisements fantasques. Tout en chantant sur le mode mineur L'amour vainqueur et la vie opportune, Ils n'ont pas l'air de croire à leur bonheur Et leur chanson se mêle au clair de lune, Au calme clair de lune triste et beau, Qui fait rêver les oiseaux dans les arbres Et sangloter d'extase les jets d'eau, Les grands jets d'eau sveltes parmi les marbres.
(English)
Moonlight Your soul is a select landscape Where charming masqueraders and bergamaskers go Playing the lute and dancing and almost Sad beneath their fantastic disguises. All sing in a minor key Of victorious love and the opportune life, They do not seem to believe in their happiness And their song mingles with the moonlight, With the still moonlight, sad and beautiful, That sets the birds dreaming in the trees And the fountains sobbing in ecstasy, The tall slender fountains among marble statues.
Paul Verlaine, 1869
(as provided by Blackrose on discord first brought up by Stardrop on discord)
—--------------------------------------------------------------------------------------
🟫 Static Frames Analysis
LIVESTREAM 4 SCHIZO
Makedonski:
Processed 18000 frames (first 10 min). Almost all frames are unique, which means that the noise is generated procedurally(TODO: check more frames, maybe i will found larger loop). Also we have some identical frames. Interestingly, they always come in pairs of 2 and there are about 1000 frames between pairs(each pair is different from another):
frame_00000151.png frame_00000150.png
frame_00001189.png frame_00001188.png
frame_00002215.png frame_00002214.png
frame_00003206.png frame_00003205.png
frame_00004312.png frame_00004311.png
frame_00005338.png frame_00005337.png
frame_00006491.png frame_00006490.png
frame_00007504.png frame_00007503.png
frame_00008516.png frame_00008515.png
frame_00009880.png frame_00009879.png
frame_00010850.png frame_00010849.png
frame_00011923.png frame_00011922.png
frame_00013104.png frame_00013103.png
frame_00014286.png frame_00014285.png
frame_00015386.png frame_00015385.png
frame_00016472.png frame_00016471.png
frame_00017478.png frame_00017477.png
Completed identical frames: 34
I checked the identical frames, found nothing but noise.
--
Looking at the frames, frame by frame, with gamma enhancement, it looks like it's just a loop of noise roughly 5 frames in length (but not exactly, sometimes it skips). The randomness is introduced due to compression artifacts and the low brightness of the noise pattern. You can see more of what the original pattern looks like by extracting the key frames only
–
Makedonski:
I don’t know exactly how compression works on YouTube, so I can’t really comment on the statement above.
I can only say that a repeating pattern does seem to be present, but only approximately.
I’ll simply provide a list of facts after studying a one-hour segment of the stream throughout the entire day:
Most of the following facts were derived from the video keyframes (1,600 frames), not from the entire segment (90,000 frames). The facts that were verified on the full segment are marked with !!.
What was checked:
Summary: The only truly interesting fact is the absence of the color (7, 7, 7), which once again points us toward STICK SEVEN.
NOT fun fact: Due to the noise, it is almost impossible to check something automatically. Therefore, all 20000 masks had to be checked by hand (with preprocessing)
If you have any other ideas on what could be done, feel free to message me in dm or just ping me.
I can also send all the code if you want to verify or test something yourself.
🟫 Minecraft Schizo
🟫 Cats & Cow Theories
The Facts:
The Shremple Raid:
On 25 March 2026, at 22:03, user @.tanki. on Discord reported that a warden had spawned at the Shremple. Then more wardens began to spawn as players started converging to defend the Shremple and a full raid started. Withers, ravagers and the ender dragon spawned as players defended the Shremple with the help of hundreds of iron golems.
During the raid several cats were spawned. Some died during the fighting so the exact number is unclear, but allegedly it was 32 spawned cats, and 15 survivors after the fight. Most of those and the cow were subsequently also killed by players and now only 1 survives. The cat has been named “filian's 9th live” by players with a nametag.
Cats:
Eats: Fish
Lives In: Jungles, Villages, Witches Huts
Special Ability: Scares away creepers in their proximity.
Cow:
Eats: Wheat
Lives In: Grassy Areas
Special Ability: Their milk cures status effects.
Theories:
Milk:
Milk cures the player of any status effects. This feature is most commonly used in Ocean Monuments to cure Mining Fatigue which has led some people to suspect the next clue will be in an Ocean Monument. Or perhaps that some status effect needs to be applied to the cats.
Cats:
Cats come from ocelots who live in the jungle. This may point to Jungle Temples as a point of interest. But cats can also be found in witches' huts and villages.
Since they also scare away creepers it might mean that we should creeper proof the temple.
Missing Info:
Since the cow and many cats died it will be impossible to determine any clue that relies on detailed information. Such as how many cats there were of which breed or color, and what IDs the cats had. This indicates that looking too closely at such information is likely a dead end since the ARG would have completely failed at this point without it.
🟫 Candle/pillar schizo
Candle Info | Corner Text Ordering Possibilities | ||
Candle Number | Candle Color | Corner text (interior glass) → Also placement order in time | Corner text (candle location) |
1 | blue | remember. | protect it. |
2 | white | watch it | its coming |
3 | purple | protect it. | remember. |
4 | black | its coming | watch it |
—
dualInfinities tried to see if a usable alphabet key could be extracted from the armor stand texts, as a means of translating the pillars:
‘protect it.’, ‘its coming’, ‘remember’, ‘watch it’
ABC_E_G, HI_ _ _MNOP, _RST_ _, W_, _ and _ ‘roct it.’, ‘it i’, ‘rememe', ‘tc t’
hmmm, Z, Q and K are reasonably easy substituted, but L and F are rough.
D can be subbed with T more-or-less, J usually gets dropped out of cipher grids anyways, X is just a compound of KS, Y is usually just a fancy i, but I don't think removing both U and V would leave a viable cipher alphabet even without F and L.
Conclusion: probably not
Note: It did inform me of the surprising number of feasible substitutions for many existing letters, allowing ciphers with fewer characters to work... a 20-character (?!) version of the english language seems surprisingly viable.
—
The four towers next to each candle were noticed to be 16 blocks tall and contain a seemingly-random assortment of blocks separate from the Quartz Pillar blocks the towers were primarily made of. Current theories include that these towers contain encoded data due to the 16-block height.
Newest theories: Each tower might be individually encoded? QR-Code? Quintecimal? Overlay certain towers on top of each other? Only one color at a time?
—------------------------------------------
Note that for all visual references,
Chiseled Quartz Blocks
have been replaced with
Redstone Blocks
for visibility.
—------------------------------------------
The towers have been unwrapped in various ways for everyone to work on:
Dynamic (rotate/reorder-able) Spreadsheet version:
Tower Blocks Spreadsheet by Asa (copy it to your own drive so you can edit it as you wish)
Note: In the spreadsheet, I included a row 0 which is the row of Quartz Pillar blocks above the lanterns that were hidden behind the upside down stairs ‘trim’ on the towers (without exception, these were all Quartz Pillar blocks, so might be nothing). I also noted the Quartz Bricks at the bottom of the towers since those blocks are the same as the rest of the temple structure rather than the towers, so possibly not considered ‘default blocks’ for the towers.
(image by Railgun on discord)
(image by coggers on discord)
Text Version:
(by GummiWyrm on discord)
Here it is numbered out:
0 = the common block (Quartz Pillar)
1 = the lanterns (cuz they are blue-ish)
2 = the chiseled quartz block
3 = the purple glass
4 = the black glass
01010101 01310101 01010101 01012101
00300000 00002000 00000200 00020000
00000040 30020000 00300000 00000004
00000000 00002000 03300000 00002004
00003000 00000000 00000002 00000000
00403002 00000322 40000000 30004220
00000000 00400000 00000000 00000000
20000000 00402002 00003000 00300000
03004020 20000000 00403000 00000000
00004000 00003000 00400200 00200200
00000002 00200200 02002000 00030004
00000000 00300022 00000002 02030002
00000000 00300000 00230000 02000000
00000000 00002000 02000000 00000044
00000004 00040040 00000000 20000000
00000000 00000000 00000000 00000000
Visualizer program to search for known values (by Helix475):
https://github.com/HeliX475/FilianARG_HelixCodes/tree/main/SlidingWindowV1
If you know Python or have an AI code helper, you’ll be able to manipulate the output to look for various things easier than by eye.
Additional Info:
Due to the chat in the SnackersHQ discord being nuked, I (Asa) figured I’d leave my records of the last conversation about the towers that I was a part of. Sadly, I only recorded the messages I thought were relevant to me, as I figured I’d be able to come back to the discord chat later for more context if I needed it. But now that’s impossible.
-----
GummiWyrm:
So im in shremple doing some investigation. The three block types on the pillars are black glass, purple glass, and chiseled blocks. Inside the temple there is a black glass "slot" that contains a chiseled block (and is the only slot that contains a block). The blue glass and white glass are not present in the pillars at all. My first thought was that the chiseled block represented another color, but since it's in the black slot that kinda discounts that idea. I still think we could be looking at layered data between the three block types on the pillars though. I'll try to create a grid representing the different colors.
-----
dualInfinities:
**so for future reference, the candle order is also the Y range, and both are the order you read Japanese in.**
—--
GummiWyrm:
yeah, i suspect lamps are to be ignored. Need to keep them in mind tho. They could also be an indicator in some way
-----
Hadez:
Zero Stripped Steam I used for these: 12334233442412114312344133214111312311142111423113344123331211113121131213311421114244414
Vertical extaction: NNSNNNNHSTALMELYALLAMTFFEUFAMMTTTTMLTEEETAPP
BIFID Output: EYGTFMCFCTIYEGLTCTEHAPSSPICFLTEATIMECTTCLFHF
-----
-----
Begin record of Discord Chat
-----
PainStalker:
so I did the grid from hans image and turned it into binary and the hex of it give us 0892a000 in one of the row which is like our 0x892a.bin folder from blackfile
PainStalker:
row 9 is 00001000100100101010000000000000 which mean 892A000 in hex numbers or 0892A000 <@475017810914050048>
PainStalker:
if you convert every line it give this in Hex
```
5755555D
80001010
09888001
80008109
00004400
7040108E
02008000
92202020
0892A000 <- this one matches the file 0x892A.bin in the recycle bin on blackfile-index
80009024
42402011
32000451
0200C040
80000103
21400080
00000000
```
Sarg:
to decimal this gets you: 튣럲펭ᄰ汬ꏀ鱠흀Û䧃൨둧訐
which translates to: "The sound of a bell ringing in the distance" (according to google)
NOTE: it’s a forced translation. It’s several languages, google will translate just about anything if you say it’s something.
-----
End record of Discord Chat
-----
Note by Asa:
After playing around with the spreadsheet trying to recreate the results that PainStalker got in the below conversation, I found PainStalker was using the following method:
Tower(Sides): NE(N,W,S,E), NW(W,S,E,N), SE(E,N,W,S), SW(S,E,N,W).
Default blocks as 0 and any non-default-block as 1 (so, binary.)
Convert each horizontal stack to binary, resulting in 4 8-long strings of binary.
Convert each 8-long binary string into hex and concatenate the result into a string.
Note by Euphoricfan1161:
I had tried to decode the pillars in base4, and all combinations of base3 and base2, each to ascii. In all orders of pillars as well as vertically or around each pillar including different possible offsets around each pillar (always top to bottom as well as left to right around). None of these returned usable strings, most of the time the characters were not even letters/numbers.
Note by Fog_runner:
Will share this just in case. If we try to analyse the towers separately, then we notice that they have “zero” layers in different spots. NE tower - layer 5, NW - layers 4,7,12,13, SE - layer 7, SW - layers 7,9. If we try to approach the numbers of candles semantically, then they could mean something like “1” - overworld (the first step in MC), “2” - Nether, “3” - the End, but there is no “4” (other than IRL). And on the top of that tower “3” (SE) and “4” (SW) share one empty layer on the same level (7). This could mean that SW should be added to others in some way or used as a key to separate other towers into segments. .
So I decided to analyse SE and SW separately. Transformed each layer into binary (0 = the common block, 1 = everything else) in 8 different rotation patterns (SENW, ENWS, NWSE, WSEN, WNES, SWNE, ESWN, NESW). Then I summed all the decimals derived from the binaries for each rotation individually, but did it in two parts. First sum is for layers 2 to 6, second sum is for layers 8 to 16. Then I summed these numbers from two towers (“w1” means I tried to use the sea lantern layer too):
I got (1560, 489) and (1575, 489) which are rather close to X 1664 and Z 439 of the End temple (bfi sign location).
However, I don’t have the bounds for the End temple so it’s hard to assess how practical these results are.
If this approach is correct, then the supposed yet hidden temple location in the Nether could be derived from NE+NW or NW+SW or NE+SW the same way but maybe with a different separator layer.
If I use layer 7 as a separator for NW+SW as well, then for NWSE and NESW rotations I get coords (X 1098, Z 701) and (X 1233, Z 671) accordingly.
If I use layer 9 as a separator for NW+SW (if SW is indeed the “key”), then for NWSE and NESW I get coords (X 924, Z 1009) and (X 1059, Z 1009) accordingly.
Update.
After Asa got the End temple bounds ( X 1595 to 1647, Y 32 to 56, Z 422 to 455 ) I noticed something else. If we take the ESWN rotation pattern, then we can get these coords: ( X 1626, Z 426 ). It means that layers 2 to 7 could encode Z and layers 2 to 16 together could encode X, thus we get coords that fit the bounds for the first time. But I don't really see how we were supposed to arrive at this conclusion without reverse engineering (if we were ever supposed to), so it's probably a coincidence (but I'm not sure).
But if this approach is correct, then the following coords could point to the temple in the Nether (redo of NW+SW, layer 7 as a separator, for all rotations with the new method):
ESWN could be the “correct” rotation because it was the one that yielded the first fitting result.
Layer 9 as a separator (same method):
All 8 variations (with X and Z being negative) for ESWN rotation for layer 7 and 9 separators yielded nothing in the Nether (checked by rixir_levi).
(X 1751, Z 389) Checked.
(X -1751, Z -389) Checked.
(X -1751, Z 389) Checked.
(X 1751, Z -389) Checked.
(X 1751, Z 721) Checked.
(X -1751, Z -721) Checked.
(X -1751, Z 721) Checked.
(X 1751, Z -721) Checked.
🟫 IRL Coords Conversion
Fog_runner: made progress on reverse engineering the temple location. If I take the exact coordinates someone wrote in the doc (1617 / 30 / 439), translate all of that into binary (11001010001 / 11110 / 110110111), then rewrite that into morse (..--.-.---. /....- /..-..-... ; 1 as ‘’.’’, 0 as ‘’-’’) then the second and third parts can be transcribed into SET (.../ ./ -) / FREE (..-./ .-./ ./ .), but the first part doesn't work like that. Maybe the ''correct'' coordinates for X are something different
Maybe we need combinations of 3 words that are somehow relevant to the arg (like (?) / SET / FREE). Then the word needs to be translated into morse, then into binary, then into decimal for actual coordinates for the minecraft server.
Altar coords reverse engineering
Then I attempted to do the same with the altar. Coords - (-401 / 101 / -3075). In binary - (110010001 / 1100101 / 110000000011). In morse, without the minuses - (..--.---. / ..--.-. / ..--------..). Tried to keep the minus in the sequence, but X only translates into a coherent word without it - MUST. Y and Z don’t yield anything, possible that they are not exactly “correct”.
So if I try Y = 102 then I get the following words - MEAN, GATE, MUTE, MIME. If I try Y = 103 it results in MUM or MUTT which don’t work at all. The word GATE at least is relevant to death and resurrection. Y = 100 results in MINE, GENE, MERE, MID. Y = 99 results in GUT.
If then I try different X values, X = -397 (with a minus) yields PEAK, ANIMA (the soul).
For clarity: - 397 = - 110001101 = - ..---..-. = ANIMA or PEAK (as coherent somewhat relevant words, ANIMA especially).
Result - (ANIMA / GATE / ?).
This is proof of concept, I didn’t even expect it would work at all. Could be useful.
Take 3 or 2 words that seem relevant together and apply the logic above (word to morse to binary to decimal) and look where it leads. Try synonyms too (like soul and anima).
TH3EEuphoricFan1161: As of Apr 9, 2026, I wrote a script to confirm the above results by Fog_runner. You can find it here: https://github.com/EuphoricFan1161/filianarg_scripts in the “Coords_and_Words” sub dir. It uses the Natural Language Toolkit (NLTK) and a small supplemental list of words (mostly abbreviations of weekdays and months). Please feel free to try it yourself and make improvements.
With this script I found that 1617 in the coordinates (1617, 30, 439) (see above) can result in the word “UNTAME”, thus giving (UNTAME, SET, FREE) as a possible set of words.
However, I have found by searching other coordinates that many sets of words may result, seemingly at random. It is likely there are many false positives using this approach…
Also of note is the convention which to define 0 := ‘-’ and 1 := ‘.’, the opposite assignment is possible as well. There is also the possibility the leading negative should be prepended to the morse string. All options may be checked in the script with the -fc and -n options. The reverse engineering of the temple location and altar coordinates by Fog_runner use different bit assignments, but are consistent in the leading negative (since the temple coordinates have no negative values)
A special thanks to Tyler/@tylerginn_ in the Snackers HQ discord for trying out this code. Without them I would not have spent more time on this script and would not have found/fixed a bug in the -n option. If you find any potential bugs please let me know :3
Fog_runner (2.05.2026): just found out you did this, thanks a lot!!!
Implementation of word-to-coord method. First two coords were the results of reverse engineering (for in depth logic look at “coords calculation” section), others are new .
Coords do not tell whether it’s overworld, nether or the end. The order is X / Y / Z.
If you want to do it yourself, a quick way to do it is to have two tabs opened (https://www.dcode.fr/morse-code ; https://www.dcode.fr/binary-code ).
Important. If word-to-binary result starts with “-” then I treat it as a minus to the decimal result (because first zero doesn’t translate into anything when you do binary-to-decimal). Only then can we get results like ANIMA and ABSENT which give negative coords close to the actual altar of resurrection.
As of yet I don’t know how to treat words like MY ( -- -.--). Is it just (-100) which in decimal is (-4)? The number of minuses is odd so it would make sense. But if the number of minuses is even do they cancel out in the context of morse-to-decimal? Left both variants just in case.
OW - overworld
N - nether
E - the end
Words | Morse | Binary | Decimal | Result |
Based on reverse engineering (proof of concept): | ||||
? / SET / FREE | (?) / ....- / ..-..-... | (?) / 11110 / 110110111 | (?) / 30 / 439 (actual coords were 1617 / 30 / 439) | E - Filian temple location in The End |
ANIMA / GATE / (ABSENT?) | -..---..-. / ..--... / - ..-------.-. | - 110001101 / 1100111 / - 110000000101 | -397 / 103 / - 3077 (actual coord were -401 / 101 / -3075) | OW - Altar of resurrection |
Ideas for coords based on the examples gathered: | ||||
LITTLE/ . / DIOMEDE | .-.. .. - - .-.. ./ . / -.. .. --- -- . -.. . | 1011110010111 / 1 / - 11110000010111 | 6039 / 1 / - 15383 | OW - nothing of interest - checked by zankjeff N - unknown E - unknown |
SAINT / . / LAWRENCE | ... .- .. -. - / . / .-.. .- .-- .-. . -. -.-. . | 1111011010 / 1 / 10111010010110101011 | 986 / 1 / 763307 | OW - unknown N - unknown E - unknown |
BERING /. / STRAIGHT | -... . .-. .. -. --. / . /... - .-. .- .. --. .... - | -11111011101001 / 1 / 1110101101100111110 | -16105 / . / 482110 | OW - unknown N - unknown E - unknown |
FIND / . / IT | ..-. .. -. -.. / . / .. - | 11011101011 / 1 / 110 | 1771 / 1 / 6 | OW - nothing of interest - checked by zankjeff N - unknown E - unknown |
FIND / THE / TRUTH | ..-...-.-.. / - ..... / - .-. ..- - .... | 11011101011 / - 11111 / - 10111001111 | 1771 / - 31 / - 1487 | OW - unknown Micahr238 checked 1771 / 63 / 1487 in the OW, wrong coords because the minus idea wasn’t flushed out yet, but it was an ocean) N - unknown E - unknown |
THE / _ / TRUTH | - ..... / ..--.- / - .-. ..- - .... | - 11111 / 110010 / - 10111001111 | - 31 / 50 / - 1487 | OW - nothing of interest (mineshaft and a trial chamber) - checked by zankjeff N - unknown E - unknown |
KEEP / . / GOING | -.- . . .--. / . / --. --- .. -. --. | - 10111001 / 1 / - - 10001101001 | - 185 / 1 / 1129 (or - 1129) | OW (both negative and positive variant of “Z”) - nothing of interest (normal spawns) - checked by zankjeff N - unknown E - unknown |
TIME / IS / UP | - .. -- . / .. … / ..- .--. | - 11001 / 11111 / 1101001 | - 25 / 31 / 105 | |
Possible but insane variants: | ||||
IT / NOT / FOUND | .. - / -. --- - / ..-. --- ..- -. -.. | 110 / - 10000 / 110100011001011 | 6 / - 16 / 26827 | OW - nothing (checked by molik0368) N - nothing (checked by molik0368) E - nothing (checked by molik0368) |
BLACK / FILE / INDEX | -... .-.. .- -.-. -.- / ..-. .. .-.. . / .. -. -.. . -..- | - 1111011100101010 / 11011110111/ 110101110110 | - 63274 / 1783 / 3446 | |
FILIAN / IS / LOST | ..-. .. .-.. .. .- -. / .. … /.-.. --- ... - | 1101111011111001 / 11111 / 10110001110 | 57081 / 31 / 1422 | N - nothing (checked by molik0368) E - nothing (checked by molik0368) |
Meme variants: | ||||
SAY / MY / NAME | ... .- -.-- / -- -.-- / -. .- -- . | 111100100 / - - - 100 / - 110001 | 484 / 4 (or -4) / -49 | |
We can apply this method in reverse if we have the coordinates (like in the reverse engineering part).
Example for the Diomede Islands coordinates (65°47′N 169°01′W):
If I take "just the whole numbers" (65 and 169), "63 times raise the parallel" (65 becomes 128), "just twice the long" (add 2 to the longitude - 169 becomes 171) I get 128 and 171. Tried to apply my method of turning decimal numbers into words, and the only fully coherent word I got out of these two is "SEAWARD". I got it by combining them (128171), turning the result into binary (11111010010101011), then into morse (.....-.--.-.-.-..). The only way to split this sequence fully into a single word is SEAWARD (... . .- .-- .- .-. -..). Tried this method to a bunch of other possible combinations (got SINGKEK once), but got nothing else of notice. Tried SEAWARD as a password for CITF_MEMO.TIF but nothing (unless capitalisation matters)
Tried to apply this method to numbers separately, but it got me nowhere. 128 171 | 10000000 10101011 | .------- .-.-.-.. | which can be transformed into “RCE”; “1ORCE” and other things, none of which worked for CITF_MEMO.TIF, but I didn’t always check different capitalisations.
2.05.2026 update
10:56 (server set time).
1056 (full) | 010000100000 | -.----.----- | YQOT
10 56 (divided into 2 parts) | 1010 00111000 | .-.- --...--- | R8J
22:56
2256 (full) | 100011010000 | .---..-.---- | JFOT
22 56 (separate) | 00010110 00111000 | ---.-..- --...--- | OL8J
🟫 Blackfile Index Schizo
🟫 Potential reverse riddle engineering
Authored and method by @TH3EEuphoricFan1161 in the Snackers HQ discord.
(Disclaimer: This solution was found from the riddle to the shremple, it was not reverse engineered. I am adding this here because many have gotten the wrong interpretation)
Found at https://www.blackfile-index.org/README.md and a small history can be found in the blackfile-index.org tab. The riddle:
0kind
It=stillnotfound
Sacrifice.complete=true
Initializing message...
“If I’m lost, the backup coordinates can be found on Earth near where east meets west: An ancient sea bridge, an island between B and C, bodies both. I have a yankee sixth sense about it. Two ivory towns. Carry the cross into the Mine, just the whole numbers, sixty three times raise the parallel, just twice the long.”
“If I’m lost, the backup coordinates can be found on Earth near where east meets west: An ancient sea bridge, an island between B and C, bodies both. I have a yankee sixth sense about it.”
→ Likely means St. Lawrence Island
“Two ivory towns.”
→ There exist two towns on St. Lawrence Island, Gambell and Savoonga
“Carry the cross into the Mine, just the whole numbers, sixty three times raise the parallel, just twice the long.”
→ The idea here is that the two towns give coordinates, somehow we should take the cross product of these, and convert the result to Minecraft coordinates. The latter parts give instructions on how.
Step 1) Define both a Cartesian and spherical coordinate system that we center the Earth on
Step 2) Convert the longitude and latitude of each town into a vector in the spherical coordinate system
Step 3) Convert the vector for each town into the Cartesian coordinate system
Step 4) Calculate the cross product of these two vectors (also calculate the reverse ordering)
Step 5) Take the resulting vector in the Cartesian coordinate system and convert back to the spherical coordinate system
Step 6) Extract the latitude and longitude from the vector in spherical coordinates
Step 7) Convert latitude to the Minecraft Z coordinate by adding 63 and rounding down for the whole number “sixty three times raise the parallel” then “just the whole numbers” (Note the idea in this step is not new and has been around for a while)
Step 8) Convert longitude to the Minecraft X coordinate by squaring and rounding down for the whole number “just twice the long” then “just the whole numbers” (Note the idea in this step is not new and has been around for a while)
(see Appendix I below for discussion on other interpretations and results)
The idea of taking a cross product is not new, however, the method here is. I cannot find anyone before trying/mentioning the idea to use 3D spherical coordinates as part of the steps.
First we will define our Cartesian and spherical coordinate system as shown below in Figure 1. This choice nicely aligns latitude with phi and longitude with theta. r could be chosen, however, we will see that when extracting the angular values at the end, input r values cancel.
Figure 1. Coordinate system setup
Next, we may begin to create the vectors in spherical coordinates, one for each town denoted by the first letter of the town. The result is shown in Figure 2.
Figure 2. Input vectors in spherical coordinates
Next we can convert these vectors in Figure 2 into the Cartesian coordinate system defined in Figure 1. This step is crucial as the cross-product is not well defined in spherical coordinates. The result is shown in Figure 3.
Figure 3. Input vectors converted to Cartesian coordinates
Next we take the cross product of the two vectors. Order of the vectors matters in the cross product, I will show one ordering. The other ordering can be obtained by swapping the subscript labels in Figure 4.
Figure 4. Vector resulting from the cross product
We define x_f, y_f, and z_f to reduce clutter and note these rely only on the input angles, not r inputs. Now with the vector resulting from the cross product, we can convert back into the spherical coordinate system shown in Figure 5.
Figure 5. Conversion back to spherical coordinates
Note that the angular components theta_f and phi_f depend only on the input angles! Extracting the new latitude and longitude, shown in Figure 6, we see that they are completely independent of r inputs. We can conclude that if we only care about extracting the angular information, we can input any arbitrary r. (A special thanks to catcaughtyourtongue/@coggers5644 in the Snackers HQ discord for the discussion leading to me deriving this result) The convention used in my code (see Python Scripts section) is to use r = 1, placing this on the unit sphere.
Figure 6. Extracting latitude and longitude from the vector in spherical coordinates
Finally we can now convert the resulting latitude and longitude to Minecraft coordinates. We add 63 to the latitude then round down to find the Minecraft z-coordinate. We square the longitude then round down to find the Minecraft x-coordinate.
The coordinates I used as input latitude and longitude are shown in Figure 7. These are taken from Google Maps’ latitude and longitude coordinates for the towns.
Figure 7. Input coordinates from Gambell and Savoonga
The two resulting minecraft coordinates are (remember the cross product result in two solutions depending on the ordering of the input vectors):
See Appendix I below for info on all solutions and results
The 2nd of these solutions leads to the so-called “shremple” which “was found by Liquid/@liquidwater2 in the Snackers HQ discord on 3/17/2026 using the konami code riddle” (see the Konami Code page), a photo of which is shown below in Figure 8. Note this README riddle came before the Konami code riddle and it is very possible they both point to the same structure. If this is the case, we can at least know what this README riddle means.
Figure 8. The “shremple” a portmanteau of shrine + temple
To calculate the result, I put together a python script. I have made it publicly available in a github https://github.com/EuphoricFan1161/filianarg_scripts. The script “cmdline_cross_prod_of_spherical.py” is in the sub directory “Blackfile-Index_Countdown_Riddle”, feel free to try it yourself and check my code. A special thanks to Tyler/@tylerginn_ in the Snackers HQ discord for their suggestion to use atan2(), which cleaned up the conversion from Cartesian coordinates to spherical coordinates section a lot. I will update the readme.txt file to point back to here when I get a chance. Below in Figure 9 is an example of the terminal output from running this code with options -ic and -v.
Figure 9. Script results with -ic and -v options (and a fun Diomede reference)
Apr 9, 2026 JST: Uploaded this document to the official ARG doc, added latitude and longitude coordinates I used as input.
Apr 9, 2026 JST: Lost access to my gitlab ¯\_(ツ)_/¯ → rehosted on github. The link is updated above. Also updated all mentions to the Name/Username convention :3
Apr 9, 2026 JST: Fixed a mistake I made when writing out the equation in Figure 4 (I accidentally wrote cos where it should have been sin) The code result is unchanged and now following the equations in the new Figure 4 you can calculate the same result as the code.
Apr 19, 2026 JST: Rechecked all reasonably possible results from the cross product in nether and end. Added Appendix I to categorize all results below
Apr 21, 2026 JST: Completed double check of all coordinates in Appendix I
May 3, 2026 JST: Added disclaimer about how this solution was not found by reverse engineering. I think it may be detrimental to the broader knowledgebase if the wrong impressions are formed about how this solution came about.
There are a few ways to read the final sentence of the riddle, notably the mathematics. These could in principle be read as addition, multiplication, and taking a power. There is also the ordering of when to take the whole numbers, either before or after these operations. Finally there is how to take the whole numbers, simply rounding down seems most sensible, but I also include rounding to the nearest whole number where it is relevant. In the table below, I lay out all readings I see as feasible and their results.
Different ways to read the riddle | Resulting Minecraft Coordinates | Results X = No Structure in ANY dimension | ||
When/how to take the whole numbers | Mathematical operation on latitude | Mathematical operation on longitude | ||
After cross product | Addition (+63) | Addition (+2) | X = 181, Z = 37 | X |
X = 1, Z = 88 | X | |||
Addition (+63) | Multiplication (x2) | X = 358, Z = 37 | X | |
X = -2, Z = 88 | X | |||
Addition (+63) | Power (^2) | X = 32041, Z = 37 | OW close to Shremple <60 blocks away | |
X = 1, Z = 88 | X | |||
Multiplication (x63) | Addition (+2) | X = 181, Z = -1638 | X | |
X = 1, Z = 1575 | X | |||
Multiplication (x63) | Multiplication (x2) | X = 358, Z = -1638 | X | |
X = -2, Z = 1575 | X | |||
Multiplication (x63) | Power (^2) | X = 32041, Z = -1638 | X | |
X = 1, Z = 1575 | X | |||
Power (^63) | Addition (+2) | X = 181, | Impossible Z val | |
X = 1, Z = 1.175x10^78 | Impossible Z val | |||
Power (^63) | Multiplication (x2) | X = 358, | Impossible Z val | |
X = -2, Z = 1.175x10^78 | Impossible Z val | |||
Power (^63) | Power (^2) | X = 32041, | Impossible Z val | |
X = 1, Z = 1.175x10^78 | Impossible Z val | |||
After cross product | Addition (+63) | Addition (+2) | X = 181, Z = 38 | X |
X = 2, Z = 88 | X | |||
Addition (+63) | Multiplication (x2) | X = 358, Z = 38 | X | |
X = 0, Z = 88 | X | |||
Addition (+63) | Power (^2) | X = 32041, Z = 38 | OW close to Shremple <60 blocks away | |
X = 0, Z = 88 | X | |||
Multiplication (x63) | Addition (+2) | X = 181, Z = -1575 | X | |
X = 2, Z = 1575 | X | |||
Multiplication (x63) | Multiplication (x2) | X = 358, Z = -1575 | X | |
X = 0, Z = 1575 | X | |||
Multiplication (x63) | Power (^2) | X = 32041, Z = -1575 | X | |
X = 0, Z = 1575 | X | |||
Power (^63) | Addition (+2) | X = 181, | Impossible Z val | |
X = 2, | Impossible Z val | |||
Power (^63) | Multiplication (x2) | X = 358, | Impossible Z val | |
X = 0, | Impossible Z val | |||
Power (^63) | Power (^2) | X = 32041, | Impossible Z val | |
X = 0, | Impossible Z val | |||
After cross product | Addition (+63) | Addition (+2) | X = 181, Z = 37 | X |
X = 1, Z = 88 | X | |||
Addition (+63) | Multiplication (x2) | X = 358, Z = 37 | X | |
X = -2, Z = 88 | X | |||
Addition (+63) | Power (^2) | X = 32143, Z = 37 | OW Shremple | |
X = 0, Z = 88 | X | |||
Multiplication (x63) | Addition (+2) | X = 181, Z = -1635 | X | |
X = 1, Z = 1634 | X | |||
Multiplication | Multiplication | X = 358, Z = -1635 | X | |
X = -2, Z = 1634 | X | |||
Multiplication (x63) | Power (^2) | X = 32143, Z = -1635 | X | |
X = 0, Z = 1634 | X | |||
Power (^63) | Addition (+2) | X = 181, | Impossible Z val | |
X = 1, Z = 1.212x10^89 | Impossible Z val | |||
Power (^63) | Multiplication (x2) | X = 358, | Impossible Z val | |
X = -2, Z = 1.212x10^89 | Impossible Z val | |||
Power (^63) | Power (^2) | X = 32143, | Impossible Z val | |
X = 0, Z = 1.212x10^89 | Impossible Z val | |||
Another way to solve the riddle
(by fog_runner, 09.05.2026).
After we got info of two new IRL locations, Just Call Me 'L' suggested to use them as a part of the riddle:
“Saw that these were located in two different places of the world. 1st image was from Japan, the 2nd image was LA. Old riddle said "If i'm lost, the backup coordinates can be found on Earth", "...between bodies B and C" ...a literal human body, and a body of land on a wall. Possibly between the two places (literally on a map) could be a clue”. L posted the following image.
I decided to follow this trail mostly because the visual "cross" imagery is too strong - the LA-JP line (yellow) crossing the Emperor seamount chain (white-ish perpendicular line on the map above), which fits the part of the old riddle about "carrying the cross". Also the Emperor seamount chain is close to the international Date Line and generally fits the description of "east meets west"
Two different possibilities emerge - I either look for existing islands or submerged structures.
1. Existing islands are mostly further south along the Hawaiian part of the Hawaiian-Emperor seamount chain. If I look there, I find two candidates – 1) Pearl and Hermes atoll or 2) Midway atoll.
The first ones have a lot of white corals and the name pearl itself invokes the “ivory nature”. But the only “yankee” part about them is that they are American Northwestern Hawaiian Islands. Nothing about the “six”.
The Midway atoll has an American base on it, and also a Memorial U.S. Marines Sixth Defense Battalion. So it hits the “yankee” part and the “six”, and it is also an atoll (white coral island with white sands) which consists of two distinct above-water parts.
the problem with this way of interpretation is that it forces to go down from the “two body” line (the Japan-LA one that we have right now).
2. The underwater structure route. This allows us to stay within the Emperor Seamounts parts exactly. The chain consists of guyots which look like giant underwater mountains with flat tops (like table surfaces). As the name suggests, most of them have names of Japanese emperors, except for one which has an American name (“yankee”) – Detroit Seamount (which seems to be named after light cruiser USS Detroit for some reason). Wikipedia also has the list of emperor seamounts with additional non-Japanese names, but all of them are located further south towards Hawaii.
Several Emperor guyots have extensive white carbonate platforms formed from ancient reefs and shallow marine sediments, but the only one that fits the description (that I have found) of “two ivory towns” is the Nintoku Seamount, which has two prominent ridges, lots of coral fragments, evidence of former atoll conditions.
For all locations I multiplied the whole numbers of coords according to the riddle (63 times the N/S and 2 times the W/E).
1) Pearl and Hermes atoll. 27.8611°N 175.8529°W. 1701 and 350.
2) Midway atoll. 28°12′27″N 177°21′00″W.1764 and 354.
3) Detroit Seamount. 51°28.80′N 167°36′E. 3213 and 334.
4) Nintoku Seamount. 41°4.80′N 170°34.20′E. 2583 and 340.
Out of these four, the one that stands out the most is Detroit Seamount. If we look at Shemple bounds we see the following space: X 32116 to 32176, Y -58 to -22, Z -45 to 36, we see that 3213 and 334 easily form the X through combination: 32134. This suggests to me that the “3” was supposed to be interpreted as a point of connection («carry the cross»?).
The second variant is more cross-like visually and gives two numbers – 32134 and 34, which fit X perfectly, but to get -34 for Y it requires a bit of moon logic – that the number 334 points downwards and thus should be interpreted as a minus. Or the 34 can be taken as is and can signify the Z dimension (it can also be taken as a negative and still work for the Z bounds).
It is possible that the "B and C, bodies both" part of the riddle was supposed to be interpreted some other way originally, and it would lead us to Shremple, but it's just too polysemantic. Who knows.
Conclusion for now: we are being purposely pointed back to Shremple, so possibly there is more to it as Hadez and Euphoric suggest.
But also I have a new location, if we apply the same logic to the other coords which can be "crossed" - the Nintoku Seamount. It gives us coords X = 25840, Z = 30 or 40 (or Y as -30 or -40). Please explore once the MC is back online. Although it's not as fitting as Detroit Seamount (it has only one point of connection, one "3" in it) and thus may be wrong
The are many Emperor Seamounts (including one with the name "Suiko", but I don’t want to get baited by this), maybe other ones can point to new locations as well when transformed using the same riddle, if they have the necessary "cross" structure like Detroit Seamount coordinates.
Diomedes relative to the JP/LA - Emperor seamount chain cross:
All things were rather close to each other
Ocean map of the similar area (look up-left for the Emperor Seamounts):
🟫 broken cursor
Researched and written by Makedonski_cf
1. I took a Base64-encoded string from the website’s styles and then decoded it into binary format.
2. I learned the ICO file structure (https://en.wikipedia.org/wiki/ICO_(file_format)) and started parsing the file byte by byte.
As a result, I learned that the ICO file stores an image in BMP format and that the icon size is 32×32.
The problem was that the data in the header contradicted itself, but since this information is not critical (it is duplicated inside the BMP), I simply skipped it and proceeded to parse the BMP.
Important: ICO files contain two images — one is the actual image, and the other is a mask that determines which pixels are transparent.
3. I learned the BMP file structure (https://en.wikipedia.org/wiki/BMP_file_format) and also started parsing it byte by byte.
Here I encountered the same issue: the headers contained contradictory information.
The only confirmed value was the icon size — 32×32.
I tried to determine the beginning of the image data, but with corrupted data this is not entirely possible: images may use a color palette of varying size, or there may be no palette at all.
Therefore, I chose the minimum possible offset (according to the BMP version 3 specification) and assumed that the image data started there.
P.S. There is a problem with this approach: we cannot be sure that we selected the correct offset (a single extra pixel shifts all subsequent ones). I provided all possible offset versions (there are 32 of them).
4. A new problem appeared: I do not know the bit depth of the image pixels, since the header contained nonsense regarding this.
However, I do know the image size (32×32 image + 32×32 mask = 32×64), as well as the number of remaining bytes.
506 bytes × 8 bits / (32 × 64) ≈ 2 bits per pixel.
It does not divide evenly, apparently because 6 bytes are missing somewhere. So I added six zero bytes at the end of the image.
Now we know that the image uses 4 different colors, so I selected an arbitrary palette:
#161616
#ab4646
#8f9bf6
#f0f0f0
5. As a result, I had all the necessary information to reconstruct the file. I rebuilt it manually (https://hexed.it/) according to Wikipedia.
However, after reconstruction, it would not open in anything except GIMP. Presumably, this is because I used an outdated BMP format during reconstruction (it is shorter).
P.S. Why did I decode the image as BMP instead of ICO? According to the ICO structure, the second half of the image should be a mask, but this did not match reality.
The mask is supposed to be 1-bit and have padding bytes at the end of each row (so that it has the same row width as the original image).
Since, according to the data, this was not the case, I decided that the best solution was not to ignore the mask, but to include it as the second half of the image instead.
Preview of “cursor” image:
Archive with all possible offsets: cursor
Update by Makedonski
I've been struggling with the cursor for a few days now (I'm 75% sure there's something there), so I'll post it here.
I ran RLE on the bitmap data and got the following sequence:
53 3 5 3 5 3 52 4 4 4 4 4 52 3 5 3 5 3 53 3 5 3 5 3 53 3 5 3 5 3 53 3 5 3 5 3 53 3 5 3 5 3 53 3 5 3 5 3 53 3 5 3 5 3 53 3 5 3 5 3 53 3 5 3 5 3 53 3 5 3 5 3 53 3 5 3 5 3 53 4 4 4 4 4 53 3 5 3 5 3 76 66 48 60 48 4 24 20 46 4 26 20 46 4 26 20 46 4 26 20 46 4 26 20 46 4 26 20 46 4 26 20 46 4 26 20 46 4 26 20 46 4 26 20 46 5 25 20 46 104 48 60 44 64 52 4 48 4 48 4 54 4 46 4 54 4 46 4 54 4 46 4 54 4 46 4 54 4 46 4 54 4 46 4 54 4 46 4 54 4 46 4 54 4 48 4 44 64 16
There are obviously repeating patterns, so I replaced them with letters:
53 3 5 3 5 3 = a (sum 72)
52 4 4 4 4 4 = a1 0
52 3 5 3 5 3 = a2 -1
53 4 4 4 4 4 = a3 +1
46 4 26 20 = b (sum 96)
48 4 24 20 = b0 0
46 5 25 20 = b1 0
54 4 46 4 = c (sum 108)
52 4 48 4 = c0 0
54 4 48 4 = c1 +2
a a1 a2 a a a a a a a a a a a3 a 76 66 48 60 b0 b b b b b b b b b b1 46 104 48 60 44 64 c0 48 4 c c c c c c c c c1 44 64 16
After that, I was left with numbers that don't fit the patterns, but I don't know what to do with them.
Maybe someone will see something in this.
—----------------------------------------------------------------------------------------------------------------------------
Additional cursor info moved here to clean up the doc:
The cursor is stored as base64 and is the following:
'data:image/x-icon;base64,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'
The header designates it is a cursor. The remainder of the header has mistakes in it, primarily the cursor size and where the hotspot is (hotspot is where the computer knows where the ‘hitbox’ of the cursor is or sort of when you click what part of the cursor is used to select the item being clicked).
The broken cursor when viewed flipped and mirrored forms 43
🟥 Blackfileindex Channel
Searching for “blackfile-index” leads to a youtube channel https://www.youtube.com/@blackfileindex/shorts
There seems to be an instagram account with the same name that posts the same content.
https://www.instagram.com/black.file.index/
Caution: second youtube short had at least 1.8k views when we found it, This is very likely unrelated to the ARG. But worth investigating.
At the time of finding it (Dec 24 3:08 AM PST, reported by fog_runner) the channel had: 1 subscriber, 5 videos (all of them shorts). Short “INCIDENT REPORT #redacted” had 1 comment (“Hello niche community❤❤” from user Notbeau123 who posted it 1 day prior). The short “RESTRICTED ZONE” had 1 comment (“❤️zesdw” from user wendyjazmindeluna8630 who posted it 2 days prior).
RESTRICTED ZONE: Saturday, 20 December 2025, 15:22:42
SLEEP STUDY #redacted: Sunday, 21 December 2025, 14:38:34
INCIDENT REPORT #redacted: Monday, 22 December 2025, 11:14:49
FIELD OBSERVATION LOG #redacted: Tuesday, 23 December 2025, 9:51:21
TEMPORAL RECORD #redacted: Wednesday, 24 December 2025, 9:19:09
ARCHIVED INCIDENT #redacted: Thursday, 25 December 2025, 11:15:55
PERSONNEL LOG #redacted:
SYSTEM RESPONSE LOG #redacted: Sat, 27 Dec 2025 11:11:34
CORRELATION NOTICE #redacted:
REDACTED RECORD #redacted:
INTERNAL REVIEW NOTE #redacted:
RESTRICTED ZONE
-- Coordinates : 37.354[ ] [ ], 7.645 [ ]
-- Magnetic interference reported
-- Electronic devices malfunction
-- Headaches reported in children
-- perimeter fenced
SLEEP STUDY / SUBJECT 44
-- Polysomnography initiated
-- No sedatives administered
-- Subject entered REM in 42 seconds
-- Audio recorded during sleep
-- Transcript not included
INCIDENT REPORT
-- Location:Neurology Ward
-- Time: 03:17
-- All monitoring alarms activated
-- No patient movement recorded
-- use undetermined
FIELD OBSERVATION LOG
-- Location: Rural Monitoring Site
-- Automated sensors activated
-- No personnel on site
-- Signal duration : 11 minutes
-- Data incomplete
TEMPORAL RECORD
-- System timestamp mismatch
-- Recorded time advanced by 47 minutes
-- No system restart detected
-- All logs verified
-- Cause undetermined
ARCHIVED INCIDENT
– File Access Restricted
– Multiple sensor conflicts
– Human error ruled out
– Review committee disbanded
– No further action taken
PERSONNEL LOG
Subject failed to report
Access badge last detected on site
No exit record found
Search protocol suspended
Entry archived
SYSTEM RESPONSE LOG:
Anomaly flag confirmed
Automated countermeasures initiated
Manual override unavailable
Monitoring extended
No user notification issued
CORRELATION NOTICE
– Independent systems recorded overlap
– Timeframe alignment confirmed
– Geographic distance deemed irrelevant
– Causal link not established
– Review pending
REDACTED RECORD
– Portions of file removed
– Access level exceeded
– Original author unidentified
– Timestamp integrity compromised
– Record sealed
INTERNAL REVIEW NOTE
– Conflicting entries detected
– Multiple versions of the same record
– Checksum verification failed
– Source discrepancy unresolved
– No correction issued
Revision Notice
-Prior entries marked obsolete
-Replacement record issued
-Legacy data retained
-Synchronization incomplete
-Proceed as normal
Residual Effect Notice
-Anomaly no longer active
-Secondary Indicators persist
-Environmental variance observed
-No mitigation required
-Monitoring reduced
Observer Error Report
-Data capture irregularities noted
-Calibration within accepted limits
-Discrepancy persists
-Observer bias not excluded
-record accepted as-is
System Trace Detected
-Secondary environments affected
-Local data irregularities logged
-Cross-check incomplete
-No immediate action required
-Monitoring ongoing
Room Access Log
-Ward: Internal Medicine
-Scheduled vacancy confirmed
-Environmental shift detected
-Cause undetermined
-Re-entry permitted
Internal Review
-Environment previously marked stable
-No environmental shift recorded
-Residual variance detected
-Source: same location
-Entry accepted
Auto-correction Routine
-Previous entries re-evaluated
-Variance reclassified as noise
-No actionable anomaly present
-Monitoring frequency reduced
-Routine resumed
Status Confirmation
-Environment integrity verified
-No unresolved discrepancies
-Auto-correction successful
-Further review unnecessary
-File Closed
TRANSMISSION RESUMED
- Upload delay acknowledged
- External access interrupted
- Data backlog detected
- File integrity pending
- Monitoring reinstated
FILE STATUS UPDATE
- Previously closed entry accessed
- Access origin unresolved
- Review window reopened
- No authorization record found
- Monitoring resumed
UNOFFICIAL SIGHTING
- Reports submitted by civilian source
- Classification pending
- No response issued
- Entry stored without review
- Follow-up not authorized
ADDITIONAL REPORTS LOCATED
- Similar entries detected
- Civilian sources unverified
- Pattern recognition incomplete
- Archive load increasing
- Escalation not approved
TRANSMISSION ERROR
- Entry received incomplete
- Frame loss detected
- Audio channel unstable
- Partial storage successful
- Resubmission pending
ENTRY RESUBMITTED
- Transmission stabilized
- Previous loss confirmed
- Internal note appended
- DO NOT RELEASE FULL INDEX
- Public access remains active
ENTRY ZONE OUTSIDE AUTHORIZED MAP
- SIGNAL LOSS BEGINS AT KM 3
- DO NOT PROCEED
FILE PARTIALLY RECOVERED
- CONTENT NOT SAFE FOR RELEASE
EXTERNAL MONITORING ACTIVE
- NO INTERVENTION AUTHORIZED
CONTAINMENT PROTOCOL ACTIVE
- ACCESS RESTRICTED
FILE VERSION UNKNOWN
- PREVIOUS ENTRIES MAY BE INACCURATE
UNAUTHORIZED AUDIO COPY
- SOURCE UNVERIFIED
Observation ongoing
(no other text on this one video itself might have some text hidden inside)
SYNC VARIANCE OBSERVED
(no other test, very high pitched noise last second or two)
Coordinates 37.354[ ] [ ], 7.645 [ ]
Interpretation №1.
If we look at Wplace similar location, we get this:
https://wplace.live/?lat=37.37860874774289&lng=7.5317867780273335&zoom=12.586042469100192
Algeria, a place close to Annaba (sector 5 on Wplace). Image itself:
Number 44 is “Killian”. Short №2 talks about Subject 44. Maybe another coincidence.
The pixels of the name “Killian” were painted by user Lunr eclipse #3622068 who has the next image next to their name (according to Wplace). Their discord name seems to be “Lunreclipse.art”.
If we flip the coordinates around:
Possible relevance of time mentioned in shorts
Error 404 (Filian is lost and not found)? Error 415? Unsupported Media Type? (#schizo)
Inverting one audio channel and combining to mono gives this visual and audible pattern:
In American Morse this could spell: cio ps iie
Could be 3 different samples. Spelling roughly:
Abccbcabbccabcbcbaaaaaaaacccccbcccabbbbbbbbccaccabcacc
Or Morse: .. . --.. -. . ----- ..... ...- .. ..- .- resulting in: iezne05viua
Mixing the stereo channels down to Mono will reveal this pattern above 15khz
Mixing stereo channels reveals a pattern similar to SLEEP STUDY:
Stereo trickery again. Ticking noise at 7kHz:
Mixing to Mono destroys most ticks:
Again, ticking at 7khz. Stereo to mono mixdown destroys the ticks.
Seems to be spelling 24211 or 24111
Attempt #2 at decoding the “Temporal record” pattern.
Tried to treat the pattern similar to “6-22-22”. Then if I see one space between two brighter “dots” I treat it as a digit separator, and if there is a larger space between the “dots” I treat it as a space between two different numbers.
Applying this logic I got: 21 2 11 53 22 1.
Applying a geohash decoder to this sequence gives coordinates Lat (λ) -37.77506500482559, Lng (φ) -179.95433270931244, just water to the right of New Zealand. Wplace also yields nothing.
Applying the default polybius grid (and rearranging the sequence into 21 21 15 32 21) yields FFEMF. If I delete single digits and apply the grid to the sequence 21 11 53 22 I get FAXG.
(anyone wants to consider the use of virtual fax machines? :schizo-clap:)
Not much of anything right now.
Attempt #3 at decoding the “Temporal record” pattern.
Here is the string of symbols I get when I look at it.
---..-.---..--.-.--------.....-...---------..-..--.--- (total of 54 symbols).
( “-” as darker segments, “.” as brighter segments). If I also try to decode it with “timing” representation in cryptii.com (signal on as “.” and signal off as “-”) I get “sh x 5”. The difference from the original version is just 2 additional spaces.
So next I transformed the string into binary 000110100011001010000000011111011100000000011011001000 (“-” into “0”, “.” into “1”). Tried to separate it into several different combinations and turned the result into decimal. Decimal for the original string is 1843469210355400.
Separated into 2 strings:
000110100011001010000000011 111011100000000011011001000
Decimal: 137340915 124782280
Separated into 3 strings:
000110100011001010 000000011111011100 000000011011001000
Decimal: 26826 2012 1736
Separated into 6 strings:
000110100 011001010 000000011 111011100 000000011 011001000
Decimal: 52 202 3 476 3 200
Into 9 strings:
000110 100011 001010 000000 011111 011100 000000 011011 001000
Decimal: 6 35 10 0 31 28 0 27 8
Into 18 strings:
000 110 100 011 001 010 000 000 011 111 011 100 000 000 011 011 001 000
Decimal: 0 6 4 3 1 2 0 0 3 7 3 4 0 0 3 3 1 0
And 27:
00 01 10 10 00 11 00 10 10 00 00 00 01 11 11 01 11 00 00 00 00 01 10 11 00 10 00
Decimal: 0 1 2 2 0 3 0 2 2 0 0 0 1 3 3 1 3 0 0 0 0 1 2 3 0 2 0
Tried to apply base 26 cipher and some other variants to a few examples but wasn’t too thorough. It yields number-letter strings, maybe something can be used as a part of yt url or something. I haven’t found anything of use yet.
Tried to apply XOR Cipher to everything and got nothing.
The only other idea I have is to use a cipher key from previous steps. Maybe “THE_TRUTH” or something. But it also hasn’t yielded any obvious results yet.
Theory:
The notes/tempo/key/chord played in the background audio of Jan 30th short "UNOFFICIAL SIGHTING" are significant to the ARG, perhaps they are supposed to correlate with something unsolved, and are meant to point the ARG towards something.
-------------------------------------------------------------------------------------------------------------------------------
BPM observations:
-Beginning notes tempo 180bpm
-later notes tempo 90bpm
Could be written on sheet music as 4/4 90 bpm
-------------------------------------------------------------------------------------------------------------------------------
Frequency observations:
NOTE: For this section, I used a recording of the YT short in audacity, and high pass filtered the audio with the settings in Ref 0.1. This was because the base audio has a lot of rumbling bass notes, and they skewed the spectrum plots pictured below.
The section with lower pitched notes goes:
406hz -> G# (x8) Ref 1.1
323hz -> E (x8) Ref 1.2
302hz -> Eb (x8) Ref 1.3
270hz -> C# (x8) Ref 1.4
with spectro analysis, it seems these are not single notes, but chords instead. Additionally, many of the frequencies do not seem to line up with music notes. The frequencies are a little flatter than the corresponding note. Frustratingly, this makes the chords nearly unidentifiable. Many of the frequencies seem to be pitched between notes. Attempting to replicate a chord with true music notes does not sound like the frequency chords. The sound comes out either in the wrong pitch or tone.
The 2nd "note" is the most spectrographically dense, consisting of the frequencies shown in Fig 1.2 the full list of notes that are the closest match to the frequencies are: Eb E G Bb Eb G. This does not match any common chords, and quite frankly sounds displeasing to the ear when played back on piano. The mixed frequencies have a much more "smooth" sound.
The section with higher pitched clear notes goes:
643hz -> E (x4) Ref 2.1
606hz -> Eb (x4) Ref 2.2
481hz -> B (x3) Ref 2.3 *NOTE it kinda sounds like a 4th note begins to play, but I cant tell if it got cut in the YT Shorts looping system or if it was actually 3 notes.
Similarly to the lower note section, the frequencies for the higher notes do not seem to perfectly line up with music notes. They are all slightly off. Some are a wider range of frequencies, Most are slightly flatter than the notes they correspond to the closest. However, the last three frequencies are a little cleaner than the low note section.
-------------------------------------------------------------------------------------------------------------------------------
Conclusion:
I entered these notes and chords into a website named chord identifier, and was unable to make any combination that makes a recognizable chord or key. This is the end limit of my musical knowledge. I am at the end of my abilities for this topic. :/
-Ratwagonz
URL For chord identifier -> https://www.scales-chords.com/chordid.php
—---------------------------------------------------------------------------------------------------------------------------
REFERENCES
Ref 0.1 -> Audacity high pass filter settings
Ref 1.1 -> Note “G#”
Ref 1.2 Note “E”
Ref 1.3 Note “Eb”
Ref 1.4 Note “C#”
Ref 2.1 Note “E”
Ref 2.2 Note “Eb”
Ref 2.3 Note “B”
🟥 Shorts narrative ideas
Maybe something of use, maybe schizo. Choose yourself