Mantra Inc.

【AEA 2020 Alumni】
– Problem:
The largest problem of global comic distribution is the slow speed / high cost of translation. These problems prevent publishers from delivering their comics abroad.
Consequently, manga (Japanese comics) are currently mostly served by low-quality, unprofessional, pirated translations on sleezy sites. The pirated Manga market in the US is estimated to be $10B/year, which is hugely damaging to publishers and Manga authors.

– Product:
We are building a comic translation tool for publishers to create an appropriate pipeline that connects comic artists and readers beyond language barriers. It contains i) a machine translation engine designed for comcis and ii) a post-editing tool for translators/letters. The translation engine achieves the best accuracy in the world (Compared with Google Translate, our engine performs similar to human translation by 3.6 points BLEU [Papineni+2002]). Combined with our online post-editing tool, the translation system accelerate the speed of comic translation by 2x.

– Business model & Tractions
We charge $1000-$5000 per month for our comic translation tool.
Within a month of the launch, 20+ publishers and online comic distributors had requested a free trial, and one of them had already switched to a paid plan ($1000/month).


Nominated by

Director, Incubation and Entrepreneurship Education
The University of Tokyo
Firstly, they are focusing on an essential problem for the publishing industry of Japan. Piracy of comics has been a problem for many years. Many top-tier publishers in Japan are tackling this issue with legal actions and service improvement.

Secondly, the co-founders have domain expertise in the area of computer science. The two co-founders of Mantra both earned PhDs from the University of Tokyo in 2019. While the comic translation is an area that had been difficult to be automated, the co-founders had already made the most accurate machine translation engine for comics in the world.

I strongly believe that both the problem they are working on and the team are worthy of being nominated for AEA2020.