How they fit together#
None of these is a side project. The generator is built from the rest, and you can check that in its package manifest rather than take our word for it.
apidef reads your OpenAPI description and extracts the semantic model. The model, and every decision your project makes about itself, is an aontu document, so a contradiction is an error at generation time and not a surprise in a customer's build. jostraca writes the files, which is why you can regenerate over an edited tree and get a three-way merge instead of a clobbered file. Underneath all of it, tabnas parses everything the toolchain reads: JSON, YAML, TOML, and jsonic, the lenient JSON dialect it grew out of.
The Voxgig tools, struct, plugin, sekreto, omni and util, are the libraries inside every generated SDK: one canonical TypeScript implementation and a port per language, held in agreement by a shared test corpus rather than by discipline. That is the only way we know to ship the same behaviour in twenty-odd languages and mean it.
Seneca is the oldest, from 2010, and the origin of the habit that runs through the rest: describe what happens as messages matched by pattern. Its matcher, patrun, still sits under the Voxgig system runtime, and the system scaffold in the voxgig org still generates a Seneca backend.