Shroomie
AI-powered news made fun and habit-forming
282 followers
AI-powered news made fun and habit-forming
282 followers
Shroomie is an AI-powered news app that makes reading fun, effortless and habit-forming. Every story is bite-sized and interactive, and every day you read, you grow a mushroom garden — turning daily news into a game. AI handles everything behind the scenes, from sourcing stories and generating visuals to personalizing your feed.
sounds like fun actually. any way to add your own topic and sources?
@math_biomes Very good point! This is on our roadmap and should be available soon. Stay tuned!
Novus
Congrats on the launch, I love whimsy software! 🍄 Too many news sites spam you with ads rather than making the content experience actually delightful.
Love the asking follow ups, what other AI features do you have planned?
Shroomie
@seandotexe Hey Sean, glad you liked it!
We have a few AI-related features in mind!
Right now, our feed personalization uses a mix of traditional ML + lightweight LLM extraction. We really want to make that more powerful, perhaps allowing users to specify what they like to see and what they don't like to see in a free-form format. Kinda like Spotify's AI playlist, but a bit more guided than that.
We are thinking of ways of customizing your mushroom buddies to "grow" with your reading habits in a more personal manner. You can imagine the buddies can have tailored appearances, tags, and achievements themed with "politics enthusiast" or "science scholar".
An AI-guided "investigative deep-dive" mode, where you can discover more about a news topic, rather than just get a summary of it. This will be functionally an extension to the follow-up feature, but it will be more delightful and interactive.
Curious what you think of AI capabilities nowadays. Have you incorporated the newer models in your older products launched a few years ago? I personally have great confidence in using 2026 AI models to improve customer-facing experiences and not just coding :p.
Congrats!
Android when? ;)
Shroomie
@hsr88 Hey Krystian, it's on our radar! Personally, Kotlin is my more familiar language, so there's that.
Why only iphone?
Shroomie
@fabian_c1 Hey Fabian, it's on our radar. Hopefully, we will have resources to port to Android soon.
DeckSpeed
Congrats on the launch! Love how Shroomie focuses not just on what news to show, but how it feels to read it. The bite-sized format + interactive flow + mushroom garden habit loop makes staying informed feel much more playful and less overwhelming.
Curious how you think about personalization, source diversity, and avoiding filter bubbles as the product grows. Excited to try it!
Shroomie
@hanzhizhang0405 Glad you like the app! Yeah personalization and source diversity were hard problems for us and (still are actually). I like to think of personalization as three parts for our product:
Diversity of news subjects during content generation. We use a combination of Tavily searches, newsletter extraction, and RSS feed adapters to gather a big pool of sources. We perform some lightweight LLM extraction on these sources, to get information like category scores, recency scores, and entities referenced by each subject. After some algorithms to select and deduplicate sources, we fire the actual content generation pipeline.
During offline training, we try to build user profiles from their explicit category rankings and implicit preferences deduced from tracking events. You know, all the simplest traditional ML stuff as a v1.
During retrieval, we can use all the tags (from step 1) we have for news subjects, and each user's profile, to generate a feed that we think they both will like and also discover something new slightly out of their comfort zone.
Hope that helps. Regarding your other question, "avoiding filter bubbles", pardon me for not understand this part of the question. Do you mind clarifying? Looking forward to chatting with you more!