AI Mnemonics Updated to use DALL·E 3!

Sorry for the radio silence for a while! We’ve been busy taking care of the young ones and in my Japanese time I’ve been focusing on studying for the JLPT coming up in December.

Anyway, just wanted to make a quiet little post that I’ve updated the AI image generation to use DALL·E 3, the updated model that you may have seen in use in the Bing image generator. It’s infinitely better quality compared to the original DALL·E that we had before, so please make use of this to make and share kanji mnemonics! It’s sometimes mind-blowing seeing the level of quality you’re able to generate.

In case you didn’t know, you can create your own mnemonics by clicking on Create on any Kanji page:

Once you’re in the Mnemonic Studio, type out a memorable story that incorporates the components and kanji meaning, making sure to surround the component names with single-#es and the kanji meaning with double-##es (as shown in the screenshot below).

Lastly, be sure to type in a good image prompt and click generate image! This will create an image based on your prompt using Open API’s DALL·E 3 image generator. When you’re done, click Publish Mnemonic and you’re all set!

There are no limits on this at the moment as we’d love to see uchisen full of mnemonics created by our users, so please generate to your heart’s content! Although please refrain from using it for non-mnemonic purposes, we have to pay for each image that’s generated :wink: