Dansday
I originally built GO BLOX because my small Indonesian Roblox community desperately needed a way to stay organized. When we were just a handful of friends, every new member brought a flood of questions and a chaotic channel layout. Relying on endless text commands was causing too much noise, and I kept seeing users make the exact same mistakes over and over. I wanted a smoother experience that felt welcoming from the moment someone joined, a phase of the project I detailed in my article about building for this community.
To fix the command confusion, I made visual interactivity a core principle from day one. I completely shifted away from plain text inputs and focused on providing tools that feel native to Discord. Under the hood, I used discord.js version 14 and built a modular architecture, dividing the application into independent pieces. The main goal was to keep the bot lightweight enough to run continuously on modest hardware, a technical challenge I discussed in my post about continuing the GO BLOX journey.
Projects like this are never really finished, and recently I have been focused on refining the user interface and stripping out unnecessary bloat. Starting around the version 1.2.0 release, I spent time evaluating what actually mattered to our server. Here is what I have been working on lately:
- Removing the chatbot functionality
To stay true to my goal of keeping the application lightweight and focused on server organization, I ended up entirely removing the chatbot feature. It was a necessary step to ensure the bot runs smoothly without wasting hardware resources. - Implementing booster roles and modals
I added dedicated logic for server boosters. Since visual interfaces are a core part of the bot, I built interactive modals for boosters to use. I recently spent time fixing modal bugs to make sure the interface stayed clean and responsive. - Adding detailed bot information modules
I integrated new visual modals specifically for displaying bot information, reducing the need for users to type out help commands. This ties directly back to my original goal of cutting down channel noise. - Building the AFK System
I introduced an AFK system to help community members keep track of who is currently available, which directly addresses the organizational needs of our growing server.
Building GO BLOX has been a great lesson in iterating based on how people actually use a tool. Instead of packing it with random features, I am constantly watching where users get stuck and updating the visual interface to guide them better.
Where this went, and one thing above that is no longer true
I am leaving the text above as it was written, because the project it describes really did end there, and correcting it in place would hide the more interesting part. Two things need saying.
The first is that GO BLOX is the ancestor of a much larger thing. On 4 April 2026 I rewrote it into a multi-tenant platform — one deployment running many bots across many servers, configured from a web panel — and that is where all the work since has gone. The economy, the market, the task engine, the AI and the public server pages all live there now, along with an MIT licence and a public repository. This project is the single-server, single-community version that taught me what the panel needed to do.
The second is a correction. Above, I describe removing the chatbot as a settled decision in service of keeping the bot lightweight. That removal happened on 15 October 2025. On 31 July 2026 I put an AI back in, and it is not a smaller feature than the one I took out: chat with per-member sessions, a live spoken conversation in voice channels, an on-device wake word, and tools for web search, wikis and the server's own data. Whatever the 2025 reasoning was worth, it did not hold for nine months, and describing that removal as a principle would be tidier than what actually happened.