With so many characters available on JanitorAI, plus hundreds more added every day, both established and new creators often share a single concern. How can you support new and smaller creators?
Creators have been facing this challenge for over a year, and JanitorAI’s growth has only made it more difficult. However, a solution might just be around the corner.
JanitorAI’s Hidden Gems Feature
On July 25, 2025, Shep, the lead developer of JanitorAI, announced a new, basic feature still in testing called “Hidden Gems.” If it functions as intended, it should help highlight smaller creators.
i think this will be a fun way to show more bots and ill figure out how else it could be approved with your feedback (basically as a global recommended) until we have semantic tagging done for a per user recommended view.
Currently, JanitorAI’s Hidden Gems feature highlights characters with fewer than 200 chats that have at least 500 tokens, contain a description, and use categorization tags. The developers plan to improve the feature based on user feedback.
Working Towards Personalized Recommendations
Shep’s announcement also reveals a plan to develop personalized user recommendations, a feature users have requested before.
Currently, users’ exposure to characters on JanitorAI is limited to the platform’s search function and basic sorting options. Users often turn to JanitorAI’s official subreddit or Discord server for recommendations to discover interesting characters they can roleplay with.
Also Read: JanitorAI Is Not Introducing Censorship Or Filters
Most social media and online content platforms have robust algorithms and proprietary systems that constantly deliver content users find engaging. This not only boosts retention but also connects creators with their target audience without requiring significant effort from the creators.
A personalized user recommendation feature on JanitorAI would help users find characters they are more likely to enjoy roleplaying with and assist creators on the platform in reaching their target audience more efficiently than they do now.
