the only thing that traycer has been missing is the semantic retrieval tool like augment code's context engine.. grep and other lsp tools are fine but they always rely on surface level matches, exact strings, symbols, or references, instead of actually understanding the intent or meaning behind the code...these will struggle once you hit 100k+ LOC codebase... there is this open source project that you can use locally and get best results even for million lines of codes.. hope traycer supports it natively
https://chunkhound.github.io/benchmark/