You could, but your reviews — specifically, how much they differ from what we expected — are what shape your taste profile. That’s the foundation of the entire machine learning algorithm. Removing games doesn’t tell us what you liked or disliked, so it weakens your recommendations.

It's not our data to use. Technically, pulling it with some GitHub library or scraping the site ourselves would be easy. But HowLongToBeat is a registered trademark of IGN Entertainment, and their site explicitly forbids reproducing any part of its content without permission. Since they don’t offer a public API or an integration path, there's nothing we can do about it at the moment.

Not yet. Every platform stores data differently, so creating a bridge with other formats and databases (Backloggd, HowLongToBeat, IsThereAnyDeal, etc.) requires custom work. If there's a specific site you'd like us to support, drop a comment so we can prioritize them after we've finished developing our own features!

Out of scope, for now. Tracking achievements across thousands of games and users would put a significant burden on the Steam Web API, which has a very strict request rate limit. We need to reserve those requests for more critical operations, like continuously scanning games and DLC and syncing collections without delay.

Through external means. To keep things safe and spam-free, we don’t currently support in-site messaging. That said, if someone has filled out Discord or Reddit info on their profile, you’re welcome to reach out there. And if you both follow each other, their Steam ID becomes visible so you can connect that way.

Doesn't matter! GamesGraph adapts to any scale you gravitate toward, so you're free to assign ratings however you like - as long as you're consistent over time.

Nope! It’s just a convenient social feature that lets you view their reviews and send or receive direct recommendations. So don’t worry — adding that one weird friend who always disagrees with you won’t mess up your results.

The rating system. While Steam also leverages advanced machine learning techniques, continuously training on a massive amount of data from millions of users, it can't easily tell if you've grown to dislike the direction of an MMO after 10 years or if you absolutely loved a single playthrough of a short indie adventure.

Sorry, it's awfully inefficient. The way machine learning works, finding the top recommendation in real-time based on active filters is already a computationally intensive task, and generating an entire sorted page would require even more resources. The current approach is much easier to optimize.

Not officially, but it’s easy to work with. Every account already comes with a unique access token (visible in your settings) that’s used by the browser extension we're testing, so while there’s no fully documented API yet, creating custom endpoints is straightforward on our end. If you’re building a personal tool or script, just reach out and explain exactly what you need - we’re happy to consider supporting it.