SortMeRNA is a production-grade bioinformatics CLI tool for filtering, mapping, and OTU-clustering next-generation sequencing reads against reference RNA databases (e.g., SILVA, Rfam). Its primary use case is removing ribosomal RNA (rRNA) from metatranscriptomic or total RNA datasets, but it can also align reads, produce OTU maps, and classify de novo sequences. The architecture is a multi-threaded C++ pipeline: reads are ingested via a flexible Readfeed (supporting gzipped, flat, or memory sources), aligned using a seed-and-extend approach with burst-trie indexes and Smith-Waterman (via Parasail), scored with Karlin-Altschul statistics (vendored ALP library), and output in SAM, BLAST, FASTX, or BIOM formats. A restart/resume mechanism persists progress in a RocksDB-backed key-value store, allowing interrupted runs to continue. The codebase is well-structured with clear module boundaries (indexing, alignment, reporting, threading), but carries legacy debt from vendored third-party code, manual memory management, and some unfinished refactors (deprecated SPLIT_READS path, dead code, commented blocks). It is actively maintained with CI builds across Linux x64/ARM64 and macOS, and includes an experimental C API for embedding in other applications.