AI automatically deciphers ancient documents, ushering in a new era of Japanese historical research. New materials related to the “Takeda Shingen Tomb” emerge from a citizen participation contest.

AI automatically deciphers ancient documents, ushering in a new era of Japanese historical research. New materials on “Takeda Shingen’s Tomb” presented in a public participation contest ​ Ina City Press Release: October 15, 2025 AI automatically deciphers ancient documents, ushering in a new era of historical research in Japan. New materials related to the “Takeda Shingen’s Tomb” emerge from a public participation contest “Minna de Transcription” now features AI translation of images of ancient documents into modern Japanese. The ancient document deciphering contest has reached 840,000 characters in the two months since its launch!
https://prcdn.freetls.fastly.net/release_image/163863/4/163863-4-be791854e7801dd75b8a887ee3f268d4-3000×2000.jpg

Documents from when Takato Stonemasons were commissioned to build Takeda Shingen’s tomb. This tomb still exists today. The “Second Ancient Document Deciphering Contest,” utilizing the AI ​​automatic deciphering and modern-language translation functions implemented on the citizen-participation ancient document deciphering platform “Minna de Honkoku,” achieved approximately 840,000 characters of decipherment within two months of the contest’s launch. Many interesting documents were discovered, including documents related to the construction of the “Takeda Shingen Tomb” and documents related to Ino Tadataka. This initiative, which uses AI technology to lower the barriers to historical research and solve local issues in a fun, e-sports-like way, is expected to rapidly advance research into undeciphered documents across the country. “Minna de Honkoku” Implements AI for Automatic Deciphering of Ancient Documents The online transcription platform “Minna de Honkoku” has been steadily adding new features since its reopening on August 2nd. Particularly noteworthy is the AI ​​automatic decoding feature. This uses AI to instantly decode cursive writing contained in images of ancient documents and even translate them into modern Japanese. This is the first service in Japan to translate images of ancient documents into modern Japanese on a platform, and we hope it will dramatically accelerate research into the hundreds of millions of undeciphered documents said to exist across the country. “Minna de Honkoku” is a citizen-participation transcription platform for ancient documents and historical materials, being developed primarily by members of the National Museum of Japanese History, the Kyoto University Paleoearthquake Research Group, and the Earthquake Research Institute at the University of Tokyo. It is being developed by Yuta Hashimoto, Associate Professor at the National Museum of Japanese History. Since its release in 2017, Minnade Honkoku has had over 10,800 participants transcribe over 48 million characters. Currently, 28 official projects are available, and the number of projects that users can freely create is increasing. Anyone can enjoy online transcriptions of a wide variety of ancient documents and historical records from all over the country.
*”Transcription” refers to the process of converting ancient documents written in cursive script into type so that they can be read by modern people. [Let’s Transcribe Together] https://honkoku.org/
https://prcdn.freetls.fastly.net/release_image/163863/4/163863-4-dbaca32b35d4b5a22a3cc0d1facf2d3e-1214×713.png “Minna de Honkoku” modern Japanese translation display screen Achieved 840,000 characters within 2 months! Ancient Document Deciphering Contest The “Second Ancient Document Deciphering Contest,” in which fierce competition is being waged on “Minna de Honkoku,” held its interim results on October 12th, and the rankings of the top 10 contestants were announced. Two months after the start, participants have deciphered approximately 840,000 characters. Among these, many interesting documents have been discovered, including documents from when the Takato Stonemasons were commissioned to build the “Takeda Shingen’s Tomb,” which still exists today in Iwakubo Town, Kofu City, Yamanashi Prefecture, and documents from when the surveying team led by Ino Tadataka came to survey the Takato Domain. With just two months left, we can’t wait to see the final results. The Ancient Document Deciphering Contest is a contest to see who can read the most cursive characters in ancient documents within a certain period of time. Top contenders will receive a total of 250,000 yen in reward money. The contest covers ancient documents held at Takato Town Library in Ina City, Nagano Prefecture. 853 documents are currently available, including historical documents related to the “Takato Stonemasons” who were active throughout the country. This event was planned by Kazuhiro Maeda, a local revitalization volunteer from Ina City, Nagano Prefecture, and is co-hosted by Ina City, Minnade Honkoku, and AMANÉ LLC. [Ancient Document Deciphering Contest Official Website] https://komonjo-contest.com [Location of Takeda Shingen’s grave, currently located in Iwakubo-cho, Kofu City, Yamanashi Prefecture] https://maps.app.goo.gl/zPJKQypMLahGDxjr5 ■Event Overview Event Period: August 2 – November 30, 2025 Participants: Anyone aged 18 or older with an interest in ancient documents How to participate: Enter on the official website ( https://komonjo-contest.com ) Prize money: 250,000 yen total (1st place: 100,000 yen, 2nd place: 50,000 yen, 3rd place: 30,000 yen, 4th to 10th place: 10,000 yen) Date: August 2, 2025: Second Ancient Document Deciphering Contest Starts & Everyone’s Transcription Relaunch November 30, 2025: Ancient Document Deciphering Contest Period Ends Mid-December 2025: Announcement of top winners Late January 2025: Awards event (Ina City & online)
https://prcdn.freetls.fastly.net/release_image/163863/4/163863-4-757965dad661133420babe24315b3116-1267×796.png Website for the Ancient Document Deciphering Contest. Enter here. The dawn of an era in which Japanese history will be rediscovered. AI and Citizens Unravel Hundreds of Millions of Undeciphered Documents It is said that there are hundreds of millions to billions of ancient documents and records from the Edo period throughout Japan, a number and variety unparalleled even by global standards. Despite this volume, there are very few experts who can read cursive script, and undeciphered documents are piling up across the country. Over the long periods of neglect, many documents have been disposed of as garbage or lost in fires and disasters. This project aims to solve this problem by leveraging two new powers: “AI technological innovation” and “gamification of citizen participation.” This project dramatically accelerates the deciphering of vast quantities of ancient documents and records, something that would be impossible for experts alone to achieve. By involving many ordinary citizens and allowing them to discover history through their own interests and with their own hands, Japan’s history and culture will be thoroughly unearthed, revealing a deeper, more clearly defined profile than ever before. We believe this initiative will pave the way for us to hand this country over to the next generation with confidence and pride.
https://prcdn.freetls.fastly.net/release_image/163863/4/163863-4-e657b6dbe42982a8a534d679efdbce5b-3000×2000.jpg This document contains written notes of Ino Tadataka’s talks about astronomy and Ezo when rain prevented surveying.
https://prcdn.freetls.fastly.net/release_image/163863/4/163863-4-38c2f488d0aca9c4d7503e8bfdc4effd-1365×1365.jpg In addition to the full-text transcription AI, this renewal adds a number of new features, including a user forum and the ability for users to create transcription projects. The previous system was released in 2019, so the technology it used was a generation or two out of date. Over the years of operation, we received various requests from participants, including requests for new features they wanted to add. This led to the current renewal, which involved a major upgrade and rebuilding the system from scratch with a new system and new technology. We were talking about releasing it this year, and then the idea of ​​this ancient document deciphering contest came up. So now was the time to do it, and I think it was a good opportunity in that sense.
https://prcdn.freetls.fastly.net/release_image/163863/4/163863-4-bc27fc163239353ca9dc90c453577bd4-1365×1365.jpg Takato Town, Ina City, has about 43,000 ancient documents. When I joined the JOCV, they didn’t know how to utilize them, and deciphering them had barely progressed. To solve this problem, we started digitally archiving the documents and launched an ancient document deciphering contest. Our goal is to have people from all over the country enjoy battling each other like in e-sports, while also resolving issues surrounding the documents and promoting the region. The new version of “Minna de Honkoku” features incredible AI, which will have a profound impact on fundamentally changing the way we decipher ancient documents and common sense. I feel like we’re finally ready to begin deciphering the hundreds of millions to billions of undeciphered ancient documents said to exist across Japan. I’m excited to see what Japan will look like when all of the documents are deciphered. Related links [Official website of the Komonjo Deciphering Contest] https://komonjo-contest.com [Official Ancient Document Deciphering Contest X] https://x.com/Komonjo_Contest [Let’s Transcribe Together] https://honkoku.org/ [AMANE LLC] https://amane-project.jp/ [Ina City, Nagano Prefecture] https://www.inacity.jp/ [Documents from when Takato Stonemasons were commissioned to build Takeda Shingen’s
tomb] https://app.honkoku.org/transcription/0ac046e2d0a32d9029d0232920de20eb/3/translation?from=/ [Location of Takeda Shingen’s grave, currently located in Iwakubo-cho, Kofu City, Yamanashi Prefecture]
https://maps.app.goo.gl/zPJKQypMLahGDxjr5