Philosophy
We treat D as a systems language worth investing in: fast to compile, pleasant to write, and capable of native tooling that does not lean on a second runtime when D can do the job.
- Supplement, don’t fork the language. Prefer libraries, tools, and docs that plug into the existing D world over alternate dialects or siloed stacks.
- Native when it matters; bridge when it’s smarter. Reimplement in D when control and deployment win (Asciidoctor, JSON5, tensors). Bind carefully when another ecosystem already solved the hard part (JPEG XL, Vello, OpenTUI).
- Ship with documentation. Notable projects land in the shared Antora docs hub so users can discover and learn them in one place.
- Practical over ceremonial. Prefer small, usable modules and services over grand frameworks that never leave the README stage.
Notable projects
A short list of org work worth opening first. Full catalogs live on GitHub.
- asciidoctor-d Native D Asciidoctor — fast AsciiDoc processing without Ruby.
- opentui-d D bindings for OpenTUI’s Zig terminal UI core.
- jxl-d JPEG XL for D: decode via jxl-rs, encode via libjxl.
- vibe-search CLIP-backed semantic search stack (core + desktop).
- densor Pure-D tensors and GGUF inference used by vibe-search.
- vector-store-dlang Vector store and embedding utilities in D.
- vello-d GPU-accelerated 2D vector graphics powered by Vello.
- repo-get Multi-VCS fetch library with SDL profiles (git, svn, hg, fossil…).
- json5-dlang Native JSON5 implementation for configuration-friendly data.
- dlang-gui-samples Side-by-side samples across D GUI toolkits.
Documentation hub
Project guides and API-oriented docs are published together with Antora. Start at the portal, then drill into each component.