Warp_cache – SIEVE cache in Rust for Python, 25x faster than cachetools
SIEVE cache beats LRU with one-line swap, but only matters if you're bottlenecked on cache.

Yet another Markdown compiler when remark and unified already dominate the ecosystem.
Static site developers, documentation tooling engineers, MDX users
remark · unified · mdsvex
I built Unifast, a markdown/MDX compiler written in Rust.
It focuses on a simpler compiler pipeline and, in my benchmarks, compiles around 25x faster than the usual remark/rehype-based setup.
This started as a personal project because I felt the existing markdown toolchain was powerful, but sometimes more layered and complex than I wanted for straightforward compilation.
It’s still very early, so I’d really appreciate bug reports, feedback, and feature requests.
Repo / docs: https://unifast.dev/ https://github.com/kenzo-pj/unifast
SIEVE cache beats LRU with one-line swap, but only matters if you're bottlenecked on cache.
Lazy deserialization means plugins only pay for nodes they actually visit.
Better than Super.so for static sites because it outputs local files instead of hosting.
Local-first Notion sync with WebP conversion beats Super.so for static sites.
Reverse-engineered undocumented MOBI format — builds dictionaries in 6 seconds vs 12 hours.
Native binaries for CI scripts beat Python startup time without Rust verbosity.