Hello, I have recently opened my long-term project for testing: GGDB (Good Gamer’s Database).

GGDB is a game database and discovery platform where players can import their gaming history from Steam, PlayStation, and other platforms, or manually add the games they have played.

The platform includes: 🎮 Gaming DNA analysis 🎯 Personalized game recommendation system ⭐ 6-layer game rating system (Gameplay, Story, Music, Visuals, Replayability, Bugs) 💬 Reviews and community interactions 🖼️ Game galleries and detailed metadata 👨‍💻 Cast & Crew system showcasing the credits and portfolios of professionals who worked on games

Based on the games users add and the ratings they provide, GGDB analyzes gaming preferences, builds a Gaming DNA profile, and generates personalized game recommendations.

The platform is currently in its testing phase and continues to evolve through user feedback. I would greatly appreciate it if you could take a look and share your thoughts.

https://www.ggdb.pro/

Thank you in advance.

  • faithcure@lemmy.worldOPEnglish
    0·
    12 days ago

    Great questions, let me break down how the pipeline works:

    Our first layer is deterministic — we pull core data from free APIs. Then multiple layers of cross-querying happen on the backend, and the results get merged before anything hits the site. After that, an AI agent runs research across the web (always with source attribution) and fills in the highest-confidence data.

    On top of that, there’s a report feature so users can flag incorrect info. Honestly, I’d argue our data quality is already higher than most sites in this space because of this layered approach.

    That said, the system is still in testing — nothing here is final, and it’ll settle into place and improve over time.

    On the “5 sources” point: which piece of info came from where is actually indicated as subtext on the page. But I hear you that it could be surfaced more clearly.

    On sourcing and licensing — we’re strict about this. We only use free APIs. For example, we don’t pull any data from MobyGames because their terms are restrictive; we only do read-only matching to check whether existing data lines up.

    As for licensing our own page content (the Terraria wiki example) — that’s a really good catch, and it’s not something we’ve formalized yet. Adding it to the to-do list. Genuinely valuable feedback, thank you.