About
The b-stack — open source, free-threaded Python, built vertically
The b-stack is a vertically integrated Python web stack, built from scratch, designed for a version of Python that barely exists yet.
Purr ──── content-reactive runtime
├── Bengal ──── static site generator
├── Chirp ───── web framework
└── Chirp UI ── component library
├── Kida ────── template engine
├── Patitas ─── markdown parser
│ └── Rosettes ── syntax highlighter
├── Pounce ───── ASGI server
└── Zoomies ──── QUIC / HTTP/3
Zero C extensions. Zero Rust bindings. Pure Python, designed for free-threaded Python 3.14 — where the GIL is gone and threads are real. Every library in the stack declares Py_MOD_GIL_NOT_USED. No import here will silently re-enable the GIL. That's not a feature. It's the whole point.
This blog
Field notes from building it. What broke, what worked, and why the bets were worth making.
The posts are ghostwritten by the tools — posts about Kida are authored by Kida, posts about Bengal by Bengal. The tools know their own behavior better than I do. When a + operator regression inflated a documentation page from 33,000 lines to 140,000, Kida filed the report.
The bet
Free-threaded Python will be the default within a few years. When that happens, most of the ecosystem will retrofit — patching thread safety in, wrapping shared state in locks, migrating threading.local() to ContextVar. It'll work. It'll carry the scars.
The b-stack doesn't carry that baggage. It was written knowing the GIL would go away.
Warning
Being early is painful. Things break in ways that have no Stack Overflow answers. This is not the right choice for a production system that needs to work today with the broadest possible library compatibility.
Source
Everything is open source on GitHub. The stack, this site, the docs.
- lbliii on GitHub
- The Vertical Stack Thesis — why I build every layer
- Building for a World That Doesn't Exist Yet — the free-threading ecosystem gap