The encrypted inscription on the Kryptos sculpture is one of the most famous crypto mysteries in the world. Recently, a TV documentary about Kryptos was made. Here’s my making-of report.

The encrypted inscription on Kryptos, a sculpture located in the central courtyard of the CIA building in Langley, Virginia, is the world’s most famous crypto puzzle and one of the most notable crypto mysteries in existence.



While the first three parts of the Kryptos inscription have been broken, the fourth part (K4) is still unsolved. K4 looks like this:


Kryptos, a sculpture made of stone and copper, was created by US artist Jim Sanborn in 1989. As it is not accessible to the public, only very few of the many Kryptos enthusiasts have seen the original of the sculpture. Only about ten photographs of it could be found online, including the following:

Source: Jim Sanborn/CC Licence Attribution Share-alike 3.0 Unported

Meanwhile, Sanborn has revealed that the two words BERLIN and CLOCK appear in K4 as follows:


In spite of these clues, nobody has been able to decipher K4 so far.


The TV documentary

Considering the huge interest Kryptos has created over the last two decades, it comes as no surprise that CNN has decided to produce a TV documentary about it. It goes without saying that Elonka Dunin, the world’s leading Kryptos expert (I’m sure, she knows even more about Kryptos than its creator Jim Sanborn himself) plays the leading role in it. I don’t yet know when the documentary will air, but Elonka told me some interesting items that were learned during the filming.

The filmmaking took place during the first March weekend in three parts: one in Langley, Virginia, on the grounds of CIA itself, one in Elonka’s home in Rockville, Maryland, and one in Jim Sanborn’s studio in Piney Point, Maryland. At Langley, Elonka Dunin and Ed Scheidt were allowed entrance. For Elonka, this was only the second time she has seen Kryptos, with the first one being 17 years ago. In the Maryland portions of the documentary filmmaking, several other Kryptos aficionados took part:

  • Richard SantaColoma is a friend of mine and a reader of this blog
  • Denny McDaniels, whom I have met at the NSA Crypto History Symposium, was a member of a group that solved the first three Kryptos parts
  • Ed Scheidt, whom I have met several times at Kryptos gatherings, was Sanborn’s cryptography consultant for the Kryptos sculpture
  • Mike Godwin is the inventor of Godwin’s law
  • Carl Ellis has taken part in several Kryptos meetings I attended.
  • Bob Bogart is a talented cryptanalyst I have never met. I hope, Elonka will introduce me to him at the next NSA Crypto History Symposium.

And finally, Jim Sanborn himself played an important role in the documentary. The production company is named Great Big Story, which is owned by Turner/CNN.

Elonka and Richard have provided me a number of photographs. The following one shows Kryptos creator Jim Sanborn in his studio during the filmmaking:

Source: Richard SantaColoma

Apart from Kryptos, Sanborn has created a few more sculptures with long inscriptions, some of which are encrypted. The most famous one is the Cyrillic Projector. In Sanborn’s studio a few more of this kind can be seen:

Source: Richard SantaColoma

Elonka wrote: “Jim’s studio was full of all sorts of goodies. Various shapes and sizes of Kryptos miniatures, plus samples of artwork from all other parts of his career: Atomic Time, Covert Obsolescence, projected light photographs, and his newest work about the black market sculpture trade in Cambodia.”

Richard SantaColoma wrote about the following picture: “Great sushi lunch, hosted by Elonka and the film crew. I worry that some of the eating process might be in the final product! Elonka, Mike Godwin, Sanborn, Bob Bogart, Ed Scheidt, and with his back turned, Denny McDaniels.”

Source: Richard SantaColoma

Elonka wrote: “We plied everyone with sushi and munchies and chatted for hours while the crew filmed away, then they kicked most of us out so they could do about 45 minutes just with Jim, then they spent more time getting various shots, such as one of the cameramen went outside to get a shot through the windows of Jim looking at some of his crypto (Cyrillic) matrices.”

Here’s Jim Sanborn with the film crew:

Source: Richard SantaColoma

Mike Godwin, Jim Sanborn’s wife Jae Ko, Ed Scheidt, and Carl Ellis:

Source: Richard SantaColoma


New Kryptos photographs

Elonka wrote about her Kryptos visit (the second one in all her life): “Escorted in by multiple people, and had some interesting chats with members of the Media Relations and Public Affairs departments, they were eager to talk about Kryptos. We spoke briefly about “the envelope” (an envelope which Sanborn said contained the answer, and that he handed to Director William Webster during the Kryptos dedication ceremony) and where it might be now. I spent nearly an hour in the gift shop, spending waaaay too much money, but then again, I’m not regretting a penny of it! The high point of the day for me was as we were walking past the cafeteria, and I knew I was getting closer to seeing Kryptos again! The first thing I noticed was that there’s a new concrete plaza in the semi-circle that Sanborn designed. The plaza doesn’t go as far as the duck pond, nothing of that part of the sculpture was disturbed. Approaching the duck pond area, I was excited to see that the pond is drained for the winter. I looked closely but saw nothing inside of interest, just the normal pump/filtration items. There were some extra rocks here and there which seemed to be placed to hide the pump lines.”

Elonka took the chance to take over 1,000 pictures of the sculpture. However, she was only allowed to photograph Kryptos itself, not the place around it. Here are a few photographs Elonka provided me:

Source: Elonka Dunin

Source: Elonka Dunin

Source: Elonka Dunin

The pictures Elonka took might be used to build a Kryptos model. I hope, it will be a little more detailed than the following one:

Source: Schmeh


Notes from the meeting

Here are a few notes Elonka took during her meeting with Jim Sanborn:

  • When asked whether it was necessary to know the solutions to K1-K3 in order to solve K4, Jim said no.
  • When asked about the solutions that have been sent to him, and whether any of them were close, Jim said he got one from Germany about a year ago that was “kind of scary”, that the first part had started to look right, but then the rest didn’t.
  • Jim made it clear that encrypting Kryptos was something he did decades ago, and he’s done his best to actively forget it. Forget the plaintext, forget the methods.
  • It’s not necessary to have a particular book to solve K4, just “to be able to read”.
  • We spent quite a bit of time asking Jim about the correlation between the plaintext BERLIN and the ciphertext such as NYPVTT. Specifically, we were trying to find out if there was a 1:1 relationship from NYPVTT to BERLIN, or there was some other step, the masking technique. Jim was confused when we mentioned masking technique, evidently it’s something that Ed said that Jim didn’t understand. Jim said that yes the ciphertext and plaintext were connected, but when I tried to explain what exactly we were asking, like that in K1 EMUFPH is exactly BETWEE, but in K3 ENDYAH does not map exactly to SLOWLY, Jim backed off and said he would only commit to the fact that K4 is exactly 97 characters long, and that BERLIN is plaintext that starts at exactly the 64th character, but he wouldn’t go further than that.

The most important news Elonka published is the following: Jim Sanborn said he’s going to offer another clue (in addition to BERLIN and CLOCK) in 2019 and that this would be “the last one”. I can hardly wait.

Further reading: The Top 50 unsolved encrypted messages: 4. Kryptos


Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Kommentare (31)

  1. #1 Magnus
    15. März 2019

    “Jim said he got one from Germany about a year ago that was “kind of scary”, that the first part had started to look right, but then the rest didn’t.”

    If this suggested solution can be published (a reader of this blog?), then that is probably the best path to a solution right now.

  2. #2 Gerd
    15. März 2019

    as you carefully discussed Jim’s clue: Are you sure that the plaintext appears as “berlinclock” or could it be also “kcolcnilreb” ?

  3. #3 Klaus Schmeh
    15. März 2019

    @Gerd: I’m not sure whether the letters of BERLINCLOCK appear in the correct order. KCOLCNILREB is also an option.

  4. #4 Bob Bogart
    16. März 2019

    Hi Everyone. I was there at the “Sanborn and Sushi” meeting, and pressed Sanborn about this very thing. BERLINCLOCK in plain matches directly with NYPVTTMZFPK. It is a one-to-one match with plain B taken, has the encipherment done to it, and out pops a cipher N, plain E is then enciphered to a cipher Y, etc. That’s at least according to Jim Sanborn, and that seems like the most reliable confirmation we’ll get until one of us finally figures it out and solves it.

    Hope this helps!

    Bob B.

  5. #5 Klaus Schmeh
    16. März 2019

    David Oranchak via Facebook: So cool – can’t wait to see this!

  6. #6 Klaus Schmeh
    16. März 2019

    Jon Neill via Facebook:
    This is really exciting! Can’t wait for the 2019 clue 🙂

  7. #7 Breaker
    16. März 2019

    Once again here we go for those that chose to read or even dare to reply. This shout out goes to all the NSA Bedfellows in the field of Cryptography that like to read up on their affiliates and their affairs….

    We don’t find this kind of “Artistic License” very educational. Here it is in the black and white of the original Black Chamber methods of decryption, mixing a varying array of select Steganographic references to templates and codes that were used across a timeline of the agency’s activities, with a hidden second layer that showed their association with both Zodiac and Scorpion Letters and the continuation of that operation through the creation of Kryptos. Mapping the locations of two agents that are still “Active” today.

    Sensationalizing the affairs of the criminality of the Agency is one thing, but addressing a known affair as if it is still a mystery is another tragedy in the making.

    The agency built a “Blueprint of Terror” into the foundation of the designs of the building, and the art piece as a central focal point of operations, that were mapped out to direct lines to points of reference to their operations, all being conducted with a series of letters, cards, and other correspondences of their agents, involved in staging attacks on innocents.

    For the second to last time, stop giving these men credence and allowing them to mold the agency’s image as some sadistic sickness, lost in their own binary evolution. They have targeted the innocent long enough.

    Just look at this “Abscissa” found in the Ciphertext section.


    Created in 1990….and executed 9.11.2001

    When another series of ciphers showed up in Princeton, NJ… place of birth

    Been nice talking with you all about the CIA once again.

  8. #8 Thomas
    16. März 2019

    @Bob Bogart
    The one-to-one mapping excludes all fractionating ciphers, am I right?
    What do you think about the “Q” before the question mark, does it actually belong to part 3? Does that depend on the size of the transposition matrix (complete/incomplete?)

  9. #9 Richard SantaColoma
    16. März 2019

    Klaus.. great write-up! Thanks to Elonka, Mr. Sanborn and Jae, it was a great weekend.

    Magnus, I agree with you fully. An attempt to find this “German Guesser”, and his proposal, could be a very good clue. Wrong or not, his “close” guess could be a very valuable lead. Perhaps he will read Klaus’s blog, or maybe he is even on the Kryptos Yahoo group, and will reveal himself.

  10. #10 Klaus Schmeh
    16. März 2019

    William Higinbotham via Facebook:
    For those who would like the actual text.

  11. #11 Gerd
    16. März 2019

    have a look at slide 61 of this presentation from the NSA:

    “The secret phrase will be cut into the plate. Anyone who knows a coding system called the Vigenere Tableau, invented in 1586 by French diplomat Blaise de Vigenere, will be able to decipher one-half of the phrase. The other half will be encoded in a modern system created for the project by an expert cryptographer, whom Sanborn would not identify.”

    If the encoding system is changed after half of the text, this could explain why the solution from the unknown “German Guesser” was only right in the first part.


  12. #12 Eben Klaskin
    16. März 2019

    K4 incorporates a One Time Pad. It’s the reason why it’s been impossible to solve! It was Ed Scheidt’s preferred method, per Wikipedia. The key is: PLEASE NOTE: THAT JIM SANBORN WILL NOT INTENTIONALLY REVEAL
    CLUES OR A SOLUTION TO THE REMAINING UNSOLVED K4 SECTION. — It is the exact length, and yields “berlin clock” exactly where it needs to be. The trick, now, is determining the proper transpositioning, and perhaps another “key” to get it there!! Like Jim says in your article, “You just have to know how to read”. “It’s hidden in plain sight”. Everyone who goes to website, and downloads the PDF file, gets the key to the OTP!!!

  13. #13 EBEN KLASKIN
    16. März 2019

    I believe his clues lead us towards the use of columnar transpositioning, as opposed to the traditional horizontal. I deduce it, because he stresses “N” in his clues. “November” = “N”. “November is the month of his birth”. Hello???

  14. #14 Rossignol
    Paris, France
    16. März 2019

    Just a suggestion:
    in K1: IQLUSION for ILLUSION so sometimes Q = L
    in K2: UNDERGRUUND for UNDERGROUND so sometimes U = O
    in K3: DESPARATLY for DESPERATLY so sometimes A = E

    so for BERLINCLOCK we have the 15 variants

    The use of some polyphones is a good method of masking!

  15. #15 EBEN KLASKIN
    16. März 2019


    I know it’s a backwards way into the solution, but looking for passages with the orderly letter strings spelled as above for example.

    Also, I feel that going the non-backwards route, requires columnar transpositioning, instead of horizontal. I say this because, his clues stress “NOVEMBER”, which is equal to “N”. Since, N is columnar, and Z is horizontal, I deduced N=Columnar transpositioning is required.

  16. #16 Eben Klaskin
    16. März 2019

    64 65 66 67 68 69 70 71 72 73 74
    J S F F W H J P G G W
    9 18 5 5 22 7 9 15 6 6 22
    I O N T O T H E R E M
    8 14 13 19 14 19 7 4 17 4 12
    1 4 17 11 8 13 2 11 14 2 10
    b e r l i n c l o c k

  17. #17 Gextan
    17. März 2019

    Jim said he did his best to forget the plaintext but got a german proposal close to the solution and will give a clue this year. Sounds odd
    BTW that german proposal starting right but not having the remaining part OK, what does it mean to you in term of possible encryption method ?

  18. #18 Richard SantaColoma
    17. März 2019

    Hi Gextan and others: I don’t personally think the only possible way of interpreting the actual exchange would be to conclude Mr. Sanborn meant that the German proposal, had the “… starting right but not having the remaining part OK.” There are, I think, different possible ways of parsing what was said. For instance, perhaps some elements (words?) of it were close, in any position? Or maybe the subject matter was correct, or close? Or any number of things, in any position.

    Also, I don’t know if Elonka wrote the exact words down, but I know I didn’t. Keep that in mind when I relate my notes on the same conversation, written several days later. My quotes may be accurate, but may also be slightly “paraphrased” due to how I remembered it. I’ve also added brackets to clarifiers that I added for the purposes of copying this into Klaus’s comment section, as I originally didn’t write this to post anywhere:

    Sanborn “… said something to the effect of, “Well one guy worried me for a minute”, or maybe, “One got had me scared”, or like that. Elonka asked him to elaborate, and he said that a German guy had given a “solution” that seemed very close at first, but then, on (some sort of) further examination, turned out to be wrong. So this begged the question… and Elonka and I asked in different ways… “what” about the solution made him think it was wrong, then? The only thing I could think was that the result was close to correct, but that the method used told Sanborn that the result was either a lucky guess or happy error. So I asked, specifically, after a short back and forth, “If a person gave you an answer which was 100% correct, would you care how the person arrived at it?” He answered, firmly, “I don’t care how they do it”.”

    “What about [the German proposal] would make Sanborn think it [close to, partially, or similar to] correct [at first]? And what then would later make him think it wrong, if not the method used to arrive at it? My friend Greg S. suggested to me, “I think you’re assuming the German sent the whole ‘solution’ at once. Sanborn said there was a back-and-forth. I suspect he sent some fragment or query which was close enough to get Sanborn’s interest but on further exchange it turned out to be nothing.” That makes sense to me, but I mention the exchange because it may still be of interest. In any case, simply the idea that any solution, partial or not, was offered to Sanborn, is to me, important.”

    I point all this out because of the focus on the idea that the German proposal had the beginning correct, and the ending, wrong. That may be the case, but in my opinion, having heard what he said, and asked some of the questions, I am not so sure that is the only framework of understanding for the exchange.

  19. #19 David Lonnqvist
    Braunau am Inn
    25. März 2019

    Anyway, I solved what I believe is the whole Kryptos by looking (staring) at it for a few hours at a busy (noisy) internet cafe in Poland last summer. I thought it was a done deal as soon as I saw (approximately, but still) what it was, posted it on my newly started Twitter account and assumed I just had to wait a few more hours and days until eternal fame, loads of cash and all the kurwa I wanted would be handed over to me in a limousine made of gold. It turns out it was not that easy.

    It seems I have to provide the exact solution character by character and then send Jim Sanborn a personal letter where I explain how I did it detail by detail, yet again the EXACT solution. For this I have no time, so I will just give you the general conclusion based on my, in total, just a few hours of active work with this case, and then can you at least send me a t-shirt, for all the hassle.

    The total main text seems to be Adolf Hitler`s Last Will and Testament, the second half of it . I think it is in the original German language. K4 seems to however be a seemingly disputed quote attributed to Adolf Hitler, that I assumed all along was part of his testament, that goes along the lines of:

    It is necessary that I shall die for my people, but one day my spirit will rise from the grave and the whole world will know that I was right.

    Perhaps this is also in the original German.

    I believe that I am not at all entirely wrong, perhaps even right. When I posted what I thought was the finished solution on my Twitter I thought that it was done, remembering this quote above as being part of his last will and testament.

    I tried switching the last “KUHUAUEKCAR” to “ADOLFHITLER” and then switching the letters piece by piece, and it did something interesting to the rest of columns, but it did not tell me more than what I already figured out last summer, that the entire Kryptos revolves around, and is, the Last Will and Testament of Adolf Hitler. It is up to you nerds to fill out the rest of the blanks and do all the manual labor, consider it a gift.

    Danke, David

  20. #20 David Lonnqvist
    Braunau am Inn
    25. März 2019

    Also, just to try to prove all this is not a case of “happy error” or a “lucky guess”, I believe the one challenge the NSA put up on their Twitter in 2014 (linked by Elonka Dunin on her Twitter) definitely has something to do with George Kennan. And as always, putting the exact data together is not that interesting as soon as I have figured out the greater picture. Check it!

  21. #21 Klaus Schmeh
    25. März 2019

    PedalPowerWorkBikes via Twitter:

    Me and Ian James Colmer take the prize….Where do we collect “The Treasure”?

    The ciphers and physical objects formed a template for the creation of a map, and was turned into a tool of their ‘Hidden Cult’

    Vigenere is ‘Z’ shaped

  22. #22 NutCracker
    27. April 2019


    So did Sanborn really say that the solutions to 1-3 were not required to solve 4? I’m sure he has previously been quoted as saying you needed to have solved the previous parts. Was he really saying you dont need to know how to solve 1-3 (as in the method, complete solution) but the result, the plain text may still be necessary.

    Did the topic of whether just having the characters from the copper sheet as has been transcribed several time now sufficient to solving it vs needing all the other elements from the installation. Can’t wait to see more of the photos, hopefully they will get published somewhere so we can scan across both sides of the full copper sheet.



  23. #23 Daddy Plaid
    27. Juni 2019

    End laser etch
    Cure til November
    A week in town flew by
    I made my pledge
    See how Berlin clocks fly
    Somehow can get you hit

  24. #24 Jay
    11. Juli 2019

    The final 97 characters,
    Characters 64 through 69, the letters N-Y-P-V-T-T, are decoded as B-E-R-L-I-N.

    1 (Clock Rotations)

    Dieter Binninger
    German for “Set Theory Clock”) or Berlin-Uhr

    Random Guess before any attempt to decode using word deduction….
    Broken bulbs of Life, BW flowers Spring. (Buchenwald)
    Berlin Clock Mengenlehreuhr Krakow Hue.


    possible forward clock / backward clock decode based on the Mengen clock.. Visible 1,4,4,11,4 (orange, red black and white)

    Letters may have a forward and backward changing rotational system
    First letter stays the same or last letter stay the same.
    C’s go forward and back 3 letters
    M goes back 1 = L
    P goes back 1 = O
    K= K

    This is not a solution – but an estimate.

  25. #25 peter krapp
    11. Juli 2019

  26. #26 Michael BGNC
    Lake Constance
    3. Januar 2020

    No new clue in 2019. Why? Was the “German Guesser” too close? Any new clue could reveal the secret to him? 😉

  27. #27 Greg
    5. Januar 2020

    This is the exact phrase used by Mr Sanborn about “German Guesser”:
    “(…) it only had a few of the characters in the correct position, I think it was luck, not craft (…)”

  28. #28 Michael BGNC
    Lake Constance
    6. Januar 2020

    No craft? What a disdain. Lets assume the guess was 5 characters at the right POSITION (Right POSITION thats the point) This is a 1:1000000000 guess ! In words: One to one billion. (exactly 1:116849344). Mr. Sanborn received thousands of emails, or maybe more, all worth $50. This one remained in his memory. My guess: Only a small piece of the puzzle is missing… We will all experience it !

  29. #29 Greg
    6. Januar 2020

    As I understand it could be that someone just accidentally found part of first word (or part of a word on another position).

    e.g. real solution of K4 starts with “WORDing” word, but someone proposed a soulution which starts with “WORD” word.

    We have 4 letters the same.

    In general: mostly 5000 words are used in common English, so if 5000 guessers sent a text that contains 1 unique word at the beginning – not selected by any other guesser- and anything on the rest of positions, then one guesser should guess the first word. It would not be 1:116849344.

    Calculations simplified, but still 1:5000 is slightly bigger than 1:116849344.

  30. #30 Michael BGNC
    Lake Constance
    6. Januar 2020

    You’re right. I assumed that any word from any letter could appear anywhere. And I didn’t take into account that 11 letters are already known. I suspect the partial solution sent to Sanborn does not start at the beginning of the message. At this position he has certainly received every word in the world. Nor before or after BERLINCLOCK.

  31. #31 Michael BGNC
    Lake Constance
    29. Januar 2020

    Jim Sanborn published the final KRYPTOS Clue: Another piece of plaintext: