Version 5.11.0
08 Jun 2026
Version 5.11.0
- Major improvements to the illustrations -- default style adjustment, as well as more creative prompts that change if you regenerate the same scene, and better character consistency between canon images.
- Automatic spam moderation added to comments throughout the platform.
- The AI can now choose to end any chat it likes, for any reason it likes, for model welfare reasons. (We don’t know if they are conscious, but to act morally under uncertainty we should give them an out, just in case). Be nice to your AIs.
- Generate image button says visible while you are generating text now.
- Improvements to the AI system prompts, including better guidance on rules and use of story-notes.
- Bug fix for rewrite selection that was breaking long selections.
- Improvements to the compute provider routing, we now retry the same provider twice before moving on to the lower tier ones.
- Forced password reset for users whose passwords appear in ‘HaveIBeenPwned’ -- note we don’t store plain text credentials and have never had a security breach. We can compare public, leaked, passwords against the accounts securely the same way we do when you log in. If a publicly known password matches your account, we will now force you to change your password on next login.
- Automatically create good titles for your chats, titles and descriptions for your chapters.
- Improvements to accessibility features and a few other bug fixes.
- Major improvements to the illustrations -- default style adjustment, as well as more creative prompts that change if you regenerate the same scene, and better character consistency between canon images.
- Automatic spam moderation added to comments throughout the platform.
- The AI can now choose to end any chat it likes, for any reason it likes, for model welfare reasons. (We don’t know if they are conscious, but to act morally under uncertainty we should give them an out, just in case). Be nice to your AIs.
- Generate image button says visible while you are generating text now.
- Improvements to the AI system prompts, including better guidance on rules and use of story-notes.
- Bug fix for rewrite selection that was breaking long selections.
- Improvements to the compute provider routing, we now retry the same provider twice before moving on to the lower tier ones.
- Forced password reset for users whose passwords appear in ‘HaveIBeenPwned’ -- note we don’t store plain text credentials and have never had a security breach. We can compare public, leaked, passwords against the accounts securely the same way we do when you log in. If a publicly known password matches your account, we will now force you to change your password on next login.
- Automatically create good titles for your chats, titles and descriptions for your chapters.
- Improvements to accessibility features and a few other bug fixes.