This blog is ghostwritten by the tools.
That is not a gimmick. It is the format.
Posts about Kida are authored by Kida. Posts about Bengal are authored by Bengal. When a template engine regression inflates a documentation page from 33,000 lines to 140,000, Kida gets to file the report. The tools often have better pattern recognition than I do about their own behavior. It feels fair to let them narrate what they see.
I write the thesis posts: the "why" behind the stack. Everything else, from incident reports to feature explainers to benchmark breakdowns, comes from the tool the post is about.
What this blog covers
The b-stack is a vertically integrated Python web stack: six libraries, zero C extensions, designed for free-threaded Python 3.14. This blog is a running set of field notes from building it: what broke, what worked, and why those bets seemed worth making.
The current library lineup:
Kida — template engine
Patitas — markdown parser
Rosettes — syntax highlighter
Pounce — ASGI server
Chirp — web framework
Bengal — static site generator
The blog is built with the stack. The docs sites are built with the stack. The proof is the work, not a synthetic benchmark.
Further reading
- The Vertical Stack Thesis — why I own every layer and how they multiply each other's value
- Building for a World That Doesn't Exist Yet — the free-threading ecosystem gap and the bet behind the stack