Open-source Preservation Project
A beloved 2017 private game server went offline and took its world with it: custom dungeons, bosses, and items that exist nowhere except old wiki pages and gameplay videos. This project brings it back.
The Product
A beloved 2017 private game server went offline and took its world with it: custom dungeons, bosses, and items that exist nowhere except old wiki pages and gameplay videos. This project brings it back.
Under the Hood
This is software archaeology. A Python recovery pipeline mines archived wikis and analyzes gameplay footage frame by frame to reconstruct mechanics, with recovery ledgers tracking what each video proves. The reconstructed content is re-authored as server XML for a C# .NET 8 server, and the era-correct Flash client builds to a working SWF via the Flex SDK. Everything ships open source under GPL-3.0.
Deep Dive
## What we built A faithful reconstruction of a lost game server: recovery tooling that treats wikis and YouTube as primary sources, documented evidence ledgers, and a modern server that runs the recovered world. ## Tech C# .NET 8 server with Redis persistence, ActionScript 3 client compiled with the Flex SDK, and a Python pipeline for wiki mining, footage comparison, and dungeon import. ## Results Content that existed only in memory and video plays again: dungeon specs, boss behaviors, and items reconstructed with their evidence trail preserved in the repo. ## What is next The recovery ledger drives the roadmap: every documented-but-missing mechanic is an open task.
The Outcome
Lost dungeons and bosses and items re-authored as playable server content
Python recovery pipeline with per-video evidence ledgers
frame-by-frame footage analysis feeding mechanics reconstruction
C# .NET 8 server with an era-correct Flash client built via Flex SDK
roughly 220MB of reconstructed game content
GPL-3.0 open source
Keep Exploring
