Thursday, October 31, 2024
I must admit, I’ve wanted to move some technical operations to a location where products directly on Roblox don’t have to process sensitive information requests. There are a few reasons for this, which I’m sure other developers can agree on, but a future where we can process requests more efficiently with the power of servers holds promise for products and services that provide more complex tasks.
As Nudle, my AI project, has been running off of its own backend, there’s no reason not to do the same with Kybo. The community around these products is growing at an alarming rate. New groups and communities are popping up everywhere, and to ensure the support for Kybo products is capable of withstanding future demands without the need for extensive manual updating is much of a concern. This doesn’t just benefit Kybo, but it benefits the people and groups we work with too.
Unfortunately, the name “Kybo compute” may seem cheesy at first, but in reality, it’s exactly that—A network that processes specific functions from hundreds of requests to minimise the use of On-Roblox processing with extreme security and scalability benefits. Plus, the name is a direct way to reference where all the processing is happening, so you’ll be able to refer to our privacy policy and terms of service if you ever find yourself confused about what’s happening.
As we slowly release new products (not limited to), you may see the term mentioned as a way to refer to how the product is processing any information and data.
Privacy
Believe it or not, in previous products, mostly due to the lack of any form of Infra, we used to store somewhat sensitive information directly within products, such as API keys and the use of other API endpoints in cases where it should be used in somewhat protected environments, which, in reality, we weren’t doing. I’m not saying any information is vulnerable; that’s not a concern, but the potential misuse of our authentication keys would be—Because quite frankly we don’t want to pay for your small project fees.
Bringing requests that require authentication keys and other sensitive information that is useful to Kybo directly onto our own servers allows us to better control and protect against any misuse. Since it’s our backend, we can have direct (full) control of how it’s accessed. While I won’t go directly into specific details regarding how the use of our servers prevents certain issues, there’s a wide range of applicable methods.
Accessibility, scalability & opportunity
Features we’ve only ever thought and dreamt of are now capable of being included in our assets thanks to the accessibility and power of Kybo compute. Rich, live, data-filled content is able to be provided to hundreds of active products simultaneously without breaking a sweat. While I won’t go into detail on the types of features, as they directly link to related upcoming projects, it’s exciting to know and explore the capabilities that our servers can provide.
Having access to such a tool allows Kybo and other creators to use the platform as a way to power their creations. We’ll work with our partners (and friends) to implement features within their products or services. This makes external server processing accessible to those who may require it for specific projects without the dreadful need for required skill and knowledge—All of which will be identifiable through a Kybo compute indicator.
Where to next
I’m not here to share any specific product announcements in this post, but hopefully in November, we'll see our first Kybo product powered by Kybo compute, alongside additional partner products in the future.
Transcripts are here!
For all readers who love audiobooks OR love to transcribe the text they're reading, most of the time, each blog post should include the addition of a transcript for you to listen to.