We're using DALL-E 3 to generate mnemonic images for every item on WK!

Absolutely i would be glad if you did ! :smiley: Also, thanks for having me on the server. よろしくです

1 Like

As a counterpoint to the early-stage AI people 違和感 I mentioned yesterday, this modern AI guy gave me a laugh this morning :laughing:


I’m glad you liked my picture :). His name is Kan! He kinda is a hero whose superpower is to feel all the powers of the heart (hence all the emoticons). Has helped me a lot with learning 感

1 Like

Just a heads up that sometimes (rarely) WK changes the mnemonic and even the radicals that make up a kanji. For example, this one for 北 was changed this week so the picture* doesn’t fit anymore.

  • That is, assuming the image I’m seeing for 北 is coming from this same plugin!

どう? :smiley:


Thanks for the heads up on that! It looks like they changed the whole radical composition on that entry. We’ll have to revise the image.

1 Like

I recognize that one. :slight_smile: That’s 腰 (waist).

“You’re flying a helicopter and you see a woman in the moon light touching her waist. She has both hands on her waist, like she’s Wonder Woman or something. But it’s the middle of the night. What is she doing?”

You really captured all the elements, it’s impressive, we’ll definitely want to add that one! Please go ahead and add it on the Discord server, let me know if you’re not sure how, I can help!

1 Like

I’ve already replaced some of the ones that you listed! Here is the new one for 胃 (stomach, meaning):

Here is the new one for 段 (steps, meaning):

Here are a couple more that I updated that you didn’t list. 腹 (belly, meaning, I took some liberties on this one):

息 (breath, reading そく):

I’ll finish replacing the others you mentioned as well. Please feel free to highlight any others that you find ugly or disturbing. :slight_smile:


Bump to a new year, new time to be alive :smiley:

What a curious way to improve learning. Props to you everyone helping to create this a better learning space <3


We’re still steadily working on these day-by-day, we now have 1031 items with both meaning and reading, and another 278 partially completed, 1877 images in all, covering 14% of all items! Anyone is welcome to join and help generate more. :slight_smile:

1 Like

Hi, I woud love to help add images. Making them is also a great way to reinforce my own learning. What’s the best way to do that? Just post them in the thread?

Never mind, I just figured out I needed to join the discord. So much fun!

1 Like

Oooo, this one is pretty :clap:


That’s one of the ones I generated. :smiley: I tried to imagine what a warehouse full of souls would look like for the kanji 倉 so I made it full of sparkly lights.

It’s surreal that this is just a normal thing now. The technology to make convincing images like this hasn’t even existed for a decade and it has only become popular in the past few years. I remember when people were making images for WaniKani manually, but we’re living in the future now!


Would I likely remember the Kanji better if I made the image manually?

Would I save a sit ton of time that could be better used learning more Kanji or Japanese by using an AI to generate it?


Bob Ross origin story