The Semantic Code team has significantly enhanced GitHub's language support system for code navigation by introducing CodeGen, which automates much of the previously labor-intensive pipeline, making it more resilient and cost-effective. This improvement broadens the range of programming languages supported on GitHub by simplifying the process of transforming source code into a format suitable for analysis, using a new system that auto-generates language-specific syntax datatypes. The previous approach required writing and maintaining two grammars per language, a process prone to errors and inefficiencies due to its manual nature and the challenges of aligning with tree-sitter's syntax nodes. By implementing named child nodes, generating Node Interface Files, and employing Haskell's generic metaprogramming framework, the new system enhances precision and reduces the engineering workload, enabling the addition of new languages like Java and CodeQL with precise Abstract Syntax Trees (ASTs). This transformation has not only improved the scalability and maintainability of the language support system but also opened up opportunities for community contributions and further advancements.