| 1 | 105,849 | 12,958 | pq_crypto | Native Ruby wrapper around ML-KEM, ML-DSA, and an optional hybrid ML-KEM-768+X25519 KEM... |
| 2 | 109,755 | 26,831 | async-background | A production-grade lightweight background job scheduler built on the Fiber::Scheduler i... |
| 3 | 124,920 | 18,038 | pg_pipeline | A driver-adjacent Ruby control-plane over ruby-pg/libpq. It multiplexes
independent, se... |
| 4 | 126,014 | 30,022 | static_embeddings | A small C-extension runtime for Model2Vec-style static embedding models. Models are con... |
| 5 | 128,393 | 12,958 | pq_crypto-seal | Versioned multi-recipient document encryption using PQCrypto hybrid KEM, HKDF-SHA256, a... |
| 6 | 162,399 | 30,022 | multi_compress | Unified C-extension gem for zstd, lz4, and brotli compression. One-shot, streaming, IO ... |
| 7 | 171,202 | 12,958 | pq_crypto-jwt | Ruby-only adapter that adds post-quantum ML-DSA JWS signing and AKP JWK/JWKS helpers to... |
| 8 | 181,971 | 36,893 | image_pack | Single API, vendored pure-C MozJPEG/libjpeg codec, in-memory compression and atomic fil... |
| 9 | 182,971 | 30,022 | secure_db_fields | A narrow production-oriented field encryption layer: AES-256-GCM envelope, HMAC-SHA256 ... |
| 10 | 185,102 | 58,818 | fast_curl | Parallel HTTP requests via libcurl curl_multi API. Releases GVL during I/O, compatible ... |
| 11 | 187,576 | 21,128 | tg_geometry | Defines TG::Geometry with immutable Geom parsing and constructor wrappers, expanded geo... |
| 12 | 188,269 | 58,818 | monotonic_grouper | Groups consecutive monotonic sequences in arrays into ranges. Supports any Comparable t... |
| 13 | 188,905 | 36,893 | swiss_hash | Swiss Table hash map as a Ruby C extension |
| 14 | 190,247 | 58,818 | fast_bloom_filter | Memory-efficient scalable Bloom Filter that grows dynamically. No upfront capacity need... |
| 15 | 193,330 | 30,022 | rack-libinjection | Native Ruby binding and Rack middleware for libinjection. Report-only by default: detec... |
| 16 | 193,440 | 58,818 | fast_string | Minimal set of optimized string scanning methods for high-throughput workloads like log... |