The Specifications and Philosophies of The Cyberix Network [PINNED]
Cyberix started as a private website for a small group of people. We had some game servers and not much else. At some point, staying a closed clique stopped being interesting, and we decided to open the site to the public. The forum launched publicly on June 7th, 2025.
The reasoning behind building this place, rather than using something that already existed, came from a few overlapping frustrations.
The first is the state of the internet as it exists now. Most platforms are built to extract attention and money from their users, not to facilitate anything of lasting value. The content produced on them reflects this. It is shallow, repetitive, and produced to satisfy algorithms rather than people. There is no real reward for quality or depth.
The second frustration is more specific. There has been a wave of communities and projects framing themselves as revivals of some better, earlier era of the internet. Most of them are costume parties. They try to replicate the visual style of old websites, old forums and old imageboards without understanding what made those places worth anything. The value was never in the aesthetics. Rather, it was in the people who were genuinely building things, writing things, and thinking through problems in public. Those people were REAL HUMANS who existed and put in real effort. That effort is what produced the culture being imitated. The imitation drops the effort and keeps only the look.
Cyberix was built to represent what that creative and intellectual output actually was, not what it looked like.
---
What Cyberix Is:
Cyberix is a hybrid of a forum and an imageboard. We describe it as an "imageforum". The core structure is that of a forum: boards, threads, threaded replies, long-term archival of conversations. On top of that we have incorporated imageboard elements. We believe that there are parts of imageboards that are inherently superior to traditional forums. Therefore, allowing both to co-exist maintains a familiar yet new environment while allowing more efficient ways to read, navigate and contribute to the site. The design of the forum means that the site is capable of supporting both fast and slow activity. Threads do not get deleted when they fall off the front page. They get buried. If someone finds a two-month-old thread worth replying to, they can.
The forum is supported by a set of complementary services.
We do not use Discord, Matrix, or any other centralized service for communication. We run IRC, XMPP, and Mumble servers. All three are bridged, meaning a message sent in one protocol appears in the others. This has practical value: people do not have to agree on a single chat client to talk in the same room. We have between 58 and 67 people connected to these services at any given time.
We run six game servers. The games we host were chosen based on a specific standard: they must be easy to launch, DRM-free or otherwise free from launcher dependency, and worth playing. We are not interested in hosting games that require a corporate platform to check whether you are allowed to run them. Several of our servers run modified or older versions of games that are no longer officially supported by their original distributors. Game nights are easy to coordinate because the chat infrastructure is already there.
We publish long-form articles on the site. These are original writing on topics relevant to the community: technology, online culture, and the various ideas we find worth arguing about and feel like they are representative of Cyberix.
---
Technical Philosophy:
Cyberix requires no JavaScript to use. There are no trackers, no advertisements, no CAPTCHAs. Tor exit nodes, VPNs, and datacenter IPs are not blocked. Anonymous posting is permitted.
The site was designed to function on Netscape 7.1, Opera 7.02, Internet Explorer 6.0, and Dillo 3.2.0 at both 1024x768 and 640x360. It also functions correctly on modern browsers and mobile devices. This was not a nostalgic exercise. Building to these constraints forces the site to remain lightweight and accessible on a wide range of hardware. A site that requires a modern browser with hardware acceleration to render a text forum is a site that has been built badly.
---
The Spam Problem and HOW we solved it:
Anonymous posting creates a persistent problem: anyone can post anything with no accountability. Most sites solve this with IP bans, CAPTCHAs, or JavaScript fingerprinting. All three add friction for legitimate users and fail against anyone with access to a proxy network.
We have taken a different, more exotic approach.
The first layer is a dynamic rate limiting system. Anonymous posting is free and frictionless until a flood pattern is detected. When that happens, the system automatically tightens limits on new anonymous posts while preserving the ability of existing users to continue. There is also an optional account system. Accounts do not force or require public usernames or post histories. They are antispam infrastructure. Accounts can be created through an invite tree: existing users can generate invite codes, and the account tree is logged internally. If a cluster of accounts tied to one branch turns out to be a spam operation, the entire branch can be removed at once. Account post limits are low for new accounts and increase as the account ages and posts normally. A spammer who tries to blast posts through a fresh account neutralizes themselves within a few posts.
The second layer is a lightweight vision model that scans all uploaded attachments and assigns them an NSFW score. Attachments that score above the threshold are blocked automatically. Moderators can override false positives manually. This was deployed to address a problem that affects every anonymous imageboard regardless of size: automated or coordinated distribution of illegal content. The model runs asynchronously and does not require human review before images go live. It also filters pornography and gore as a secondary effect, keeping the visual interface SFW without requiring constant human attention.
The third layer, deployed on March 8th, 2026 (one day ago as of this thread's creation), is an asynchronous LLM moderation system. It processes posts in batches and flags content that is low-quality, bad-faith, or disruptive by the standards of the site rules. It does not target posts simply for being negative or controversial. It targets posts that have no productive content and exist only to degrade the quality of a thread. Phase 1 is currently running with public warning banners on flagged posts so the community can surface false positives before the system is trusted with autonomous action. The goal is to handle the moderation problem that cannot be solved with rate limits: the human poster who is present and deliberate but contributing nothing.
---
Growth:
Cyberix is quickly growing at an exponential rate.
From June 7th to December 31st, 2025, Cyberix recorded 932,589 total hits across 392,444 unique visitors.
January 2026: 704,264 hits, 456,135 unique visitors.
February 2026: 680,561 hits, 386,573 unique visitors.
March 2026, first nine days: 166,757 hits, 100,979 unique visitors.
All-time as of this post: 2,484,171 hits, 1,336,131 unique visitors.
---
Operations:
Cyberix runs on a VPS costing 36 euros per quarter. The domain costs $15 per year. The monthly donation goal to cover costs is $14. The server infrastructure is containerized and modular, meaning we can migrate to a different provider rapidly if necessary. Nightly database backups run automatically and our server files are backed up roughly monthly.
We attempted to bring in passive income through advertising to reduce reliance on donations. We applied to EthicalAds, Carbon Ads, and BuySellAds. All three rejected us. Their requirements ranged from demanding analytics tools we do not use to restricting publishers to developer-specific niches. We are not interested in ads that require tracking users, breaking old browser support, or compromising the site's character. We are not in a position where this is a crisis. Cyberix runs cheaply and the donation goal is low.
---
Philosophy of the Network:
We want Cyberix to be a place where people produce and preserve things worth reading and using. We are looking for long threads that develop real arguments, projects that create real solutions and conversations that would not survive on a platform built around engagement metrics because they are too slow, too long, or too abrasive for the algorithm.
We are not trying to reproduce the "aesthetic" of an older internet. We are trying to produce something that has the same qualities that made such an era worth remembering: people building things, thinking in public, and not being managed by a corporate structure that profits from their activity.
The site is independent, small, and intends to stay that way.
The reasoning behind building this place, rather than using something that already existed, came from a few overlapping frustrations.
The first is the state of the internet as it exists now. Most platforms are built to extract attention and money from their users, not to facilitate anything of lasting value. The content produced on them reflects this. It is shallow, repetitive, and produced to satisfy algorithms rather than people. There is no real reward for quality or depth.
The second frustration is more specific. There has been a wave of communities and projects framing themselves as revivals of some better, earlier era of the internet. Most of them are costume parties. They try to replicate the visual style of old websites, old forums and old imageboards without understanding what made those places worth anything. The value was never in the aesthetics. Rather, it was in the people who were genuinely building things, writing things, and thinking through problems in public. Those people were REAL HUMANS who existed and put in real effort. That effort is what produced the culture being imitated. The imitation drops the effort and keeps only the look.
Cyberix was built to represent what that creative and intellectual output actually was, not what it looked like.
---
What Cyberix Is:
Cyberix is a hybrid of a forum and an imageboard. We describe it as an "imageforum". The core structure is that of a forum: boards, threads, threaded replies, long-term archival of conversations. On top of that we have incorporated imageboard elements. We believe that there are parts of imageboards that are inherently superior to traditional forums. Therefore, allowing both to co-exist maintains a familiar yet new environment while allowing more efficient ways to read, navigate and contribute to the site. The design of the forum means that the site is capable of supporting both fast and slow activity. Threads do not get deleted when they fall off the front page. They get buried. If someone finds a two-month-old thread worth replying to, they can.
The forum is supported by a set of complementary services.
We do not use Discord, Matrix, or any other centralized service for communication. We run IRC, XMPP, and Mumble servers. All three are bridged, meaning a message sent in one protocol appears in the others. This has practical value: people do not have to agree on a single chat client to talk in the same room. We have between 58 and 67 people connected to these services at any given time.
We run six game servers. The games we host were chosen based on a specific standard: they must be easy to launch, DRM-free or otherwise free from launcher dependency, and worth playing. We are not interested in hosting games that require a corporate platform to check whether you are allowed to run them. Several of our servers run modified or older versions of games that are no longer officially supported by their original distributors. Game nights are easy to coordinate because the chat infrastructure is already there.
We publish long-form articles on the site. These are original writing on topics relevant to the community: technology, online culture, and the various ideas we find worth arguing about and feel like they are representative of Cyberix.
---
Technical Philosophy:
Cyberix requires no JavaScript to use. There are no trackers, no advertisements, no CAPTCHAs. Tor exit nodes, VPNs, and datacenter IPs are not blocked. Anonymous posting is permitted.
The site was designed to function on Netscape 7.1, Opera 7.02, Internet Explorer 6.0, and Dillo 3.2.0 at both 1024x768 and 640x360. It also functions correctly on modern browsers and mobile devices. This was not a nostalgic exercise. Building to these constraints forces the site to remain lightweight and accessible on a wide range of hardware. A site that requires a modern browser with hardware acceleration to render a text forum is a site that has been built badly.
---
The Spam Problem and HOW we solved it:
Anonymous posting creates a persistent problem: anyone can post anything with no accountability. Most sites solve this with IP bans, CAPTCHAs, or JavaScript fingerprinting. All three add friction for legitimate users and fail against anyone with access to a proxy network.
We have taken a different, more exotic approach.
The first layer is a dynamic rate limiting system. Anonymous posting is free and frictionless until a flood pattern is detected. When that happens, the system automatically tightens limits on new anonymous posts while preserving the ability of existing users to continue. There is also an optional account system. Accounts do not force or require public usernames or post histories. They are antispam infrastructure. Accounts can be created through an invite tree: existing users can generate invite codes, and the account tree is logged internally. If a cluster of accounts tied to one branch turns out to be a spam operation, the entire branch can be removed at once. Account post limits are low for new accounts and increase as the account ages and posts normally. A spammer who tries to blast posts through a fresh account neutralizes themselves within a few posts.
The second layer is a lightweight vision model that scans all uploaded attachments and assigns them an NSFW score. Attachments that score above the threshold are blocked automatically. Moderators can override false positives manually. This was deployed to address a problem that affects every anonymous imageboard regardless of size: automated or coordinated distribution of illegal content. The model runs asynchronously and does not require human review before images go live. It also filters pornography and gore as a secondary effect, keeping the visual interface SFW without requiring constant human attention.
The third layer, deployed on March 8th, 2026 (one day ago as of this thread's creation), is an asynchronous LLM moderation system. It processes posts in batches and flags content that is low-quality, bad-faith, or disruptive by the standards of the site rules. It does not target posts simply for being negative or controversial. It targets posts that have no productive content and exist only to degrade the quality of a thread. Phase 1 is currently running with public warning banners on flagged posts so the community can surface false positives before the system is trusted with autonomous action. The goal is to handle the moderation problem that cannot be solved with rate limits: the human poster who is present and deliberate but contributing nothing.
---
Growth:
Cyberix is quickly growing at an exponential rate.
From June 7th to December 31st, 2025, Cyberix recorded 932,589 total hits across 392,444 unique visitors.
January 2026: 704,264 hits, 456,135 unique visitors.
February 2026: 680,561 hits, 386,573 unique visitors.
March 2026, first nine days: 166,757 hits, 100,979 unique visitors.
All-time as of this post: 2,484,171 hits, 1,336,131 unique visitors.
---
Operations:
Cyberix runs on a VPS costing 36 euros per quarter. The domain costs $15 per year. The monthly donation goal to cover costs is $14. The server infrastructure is containerized and modular, meaning we can migrate to a different provider rapidly if necessary. Nightly database backups run automatically and our server files are backed up roughly monthly.
We attempted to bring in passive income through advertising to reduce reliance on donations. We applied to EthicalAds, Carbon Ads, and BuySellAds. All three rejected us. Their requirements ranged from demanding analytics tools we do not use to restricting publishers to developer-specific niches. We are not interested in ads that require tracking users, breaking old browser support, or compromising the site's character. We are not in a position where this is a crisis. Cyberix runs cheaply and the donation goal is low.
---
Philosophy of the Network:
We want Cyberix to be a place where people produce and preserve things worth reading and using. We are looking for long threads that develop real arguments, projects that create real solutions and conversations that would not survive on a platform built around engagement metrics because they are too slow, too long, or too abrasive for the algorithm.
We are not trying to reproduce the "aesthetic" of an older internet. We are trying to produce something that has the same qualities that made such an era worth remembering: people building things, thinking in public, and not being managed by a corporate structure that profits from their activity.
The site is independent, small, and intends to stay that way.