As the co-creator of this system, the initial idea was to build a recommender focused on books, but then we thought: why stop there? Why not mix everything together—movies, TV series, anime, educational content—and see where it leads us? This approach not only broadens the scope of recommendations but also offers a unique gateway for people who may have never explored certain content types before. For instance, someone who loves movies and TV series about a particular topic is likely to enjoy related books or even dive into anime or educational content they might not have considered otherwise.

One of the most enjoyable aspects of this system is seeing how different collections come together. Sometimes, you stumble upon a piece of content in a collection and think, 'Yeah, in a really strange and unexpected way, this totally belongs here!' Those moments often lead to a lot of laughter and surprise, which is exactly the kind of reaction we hoped to inspire.

However, there are still areas that need improvement. Not all content types consistently contain the most accurate examples, so it's important to approach the collections with a bit of flexibility—seeing them as more inspirational or as closely related to the theme rather than exact matches.

Overall, it's been fascinating to see how this platform brings together content in such a unique and engaging way, and I'm excited to keep refining it to ensure it truly delivers on its potential.

As one of the creators, I'm naturally biased, but I’m genuinely pleased with how the system turned out. Here are a few highlights:

1) Unified Content Hub: Having movies, TV shows, and books all aggregated in one place is incredibly practical. It simplifies the browsing experience.
2) Quick and Relevant Results: With just a few clicks, the system reliably guides users to collections and content that align with their interests.
3) New Discoveries: I’ve found movies, shows, and books I didn’t know existed, which makes me confident that other users may find it valuable as well.
4) Powered by Our Tech: It’s particularly rewarding that the entire system was built using ChatGPT and NextCreature—our startup’s own software. It was essential for us to prove the value of our tools in a real-world application.

Areas for Improvement:

1) Content Accuracy: Not all items in the collections are perfect; some could be more relevant.
2) ChatGPT Reasoning: our tests have shown that sometimes suggested movie (or a book) seemed not that relevant to a specific collection, but if we asked ChatGPT to explain why this movie was included in this collection, we tend to agree with ChatGPT. In other words, ChatGPT Reasoning really adds value to the user experience and it’s not included in the current version
3) User Feedback Integration: I’d like to see more features that allow users to provide feedback directly, helping us refine the system further.
4) Collaboration Features: Introducing ways for users to contribute to the recommender’s evolution would add a valuable community-driven aspect.
5) Expanding Topics: There’s room to generate more collections to cover a broader range of topics, and we plan to work on that soon.

I am absolutely amazed by this recommender system! The ability to discover new content across different media is something I never knew I needed. I’m a huge fan of books, and the recommendations provided were spot-on, introducing me to titles I never would have found on my own. It's incredible how it doesn't just extract preferences but introduces you to new genres and styles effortlessly.

Well done! Keep the momentum going!

A great tool for discovering new content! The interface is sleek, and the suggestions are very relevant. I would love to see a feature where we can share our favorite recommendations with friends.

The variety of recommendations is fantastic! It's amazing how the system can suggest new shows and books that align so well with my tastes. I wish there was a way to filter out content I've already seen or read

Very pleased with the personalized suggestions across different media. It's user-friendly and efficient. Maybe consider adding a randomizer for those who are feeling adventurous!

This system has completely transformed the way I explore new content. I never thought I would be interested in anime, but the recommendations are so tailored and well-presented that I found myself diving into this new world with excitement. The way it integrates educational content related to my favorite movies is also fascinating. Truly an impressive piece of technology!

very cool project, halfway through the creation video and excited about what I might discover here, been trying out different anime recomender systems lately and you've got them all beat.

there's just one feature request:

It seems that for niche playlists (>2 tags), for media that is released in multiple languages like anime, it can occure that both the english version and version and other language (e.g. japanese) show up.

for example, when selecting the tags for anime:
"detective ", "classic" and "horror",

and clicking on:
Anime Series: Top 5 Supernatural Thrillers Involving Demonic Possession and Exorcisms

you'll get both "blue exorcist" and "ao no exorcist", which are the same series.

having a check for this in place (which could be done in a dynamic-programming-esque fashion to save compute w.r.t. media that will never be compared).

very nice work!

the platform is super convenient to use to find any source of entertainment with specific content!

I think this website is super useful for anyone who likes books, movies, shows or even anime, either it be for entertainment or for the culture. It is inspiring and helps people to enjoy new work of art and take deeper dive into their loved genres. The website is also easy to understand and use, as well the colours are very calming. I like how many possible tags and genres it is possible to look into and how specific it can get! An amazing product!