December 9, 2021

AI Dungeon 2D Beta

From the beginning of AI Dungeon we’ve dreamed of the future of living dynamic experiences built with AI. Just as games in the past started with text and moved to more immersive mediums we saw that this new evolution of experiences would go through the same evolution that games did decades ago.

We’re excited to announce the beta release of the next evolution of AI powered experiences, AI Dungeon 2D.

To get access players can subscribe to the Gold or Platinum plans. For every action you take, the AI will deliver richly detailed pixelated images within the context of your story.

Check this one out!

AI brings a scenic forest path to life.

Here’s how to turn on AI Dungeon 2D:

  1. Go to the Settings tab on the left side of the AI Dungeon app
  2. Turn on AI Dungeon 2D
  3. Open a story and start playing
  4. Gush over the pixelated art (as a true gaming aficionado would)

In case you’re wondering how detailed AI-generated pixels can get, below are more examples we’ve created in our testing:

A Jedi mouse appears!
POV: You’re a cyberpunk hacker riding a dragon in Rome.
BLAST FROM THE PAST: You meet Steve Jobs to demo your new Banana Phone idea.‌‌
A classic castle in the forest
Can’t have an adventure without enemies to face! The AI threw this evil necromancer at you.

We look forward to your feedback on this major step forward in how you can experience stories in AI Dungeon! Reach out to us anytime at [email protected] for questions and feedback.


Who will get access to AI Dungeon 2D?
To get access players can subscribe to the Gold or Platinum plans.

How do I use AI Dungeon 2D?
Settings → AI Dungeon 2D, then play one of your stories.

What model are you using?
We continue to explore additional models, but this pixel model uses Pixray by dribnet. Please contact [email protected] before making NFTs using generated images.

Are the images generated unique each time?
Right now the most relevant pre-generated images from a cache (which we are continuing to grow) are matched to users stories which allows us to deliver images quickly and cheaply. We're also working on adding a manual generation mode which will take longer and cost scales, but allow users to generate completely unique images.

What else is on the roadmap?
We're continuing to improve the quality and size of the pixel images, as well as working on new even better styles. We'll also be adding ways for users to report poor / inappropriate images so we can improve the cache.

Can I generate sexual content?
We currently don't allow the AI to generate sexual / nudity image content. That includes images that are in the cache and delivered to users' stories.