Key information
Capcom has released its data display and serialization foundation library, RE:Dox, as open-source software under the Apache-2.0 license. Serving as foundational technology for the next-generation engine project REX, RE:Dox introduces AI and machine learning integration into the RE ENGINE development workflow.
RE:Dox functions as a .NET Token-based structured data engine. It converts various datasets into unified formats to accelerate reading, editing, and processing tasks. Official benchmarks show that parallel JSON deserialization can achieve speeds up to approximately 2.84 times faster than standard .NET System.Text.Json.
The REX project consists of six distinct initiatives designed to enhance the RE ENGINE. These include RE:UI for optimizing developer tools with AI, RE:Runtime for grouping object animations to reduce burdens, and RE:Flows for simplifying gameplay mechanic implementation through automation rather than generative content.
Capcom clarified that AI is utilized strictly to improve productivity and tool efficiency rather than directly generate artistic assets. Over 2,000 developers worldwide currently utilize the RE ENGINE.
