Go
stable
schemaflux
Compile structured data from anything.
go install github.com/greynewell/schemaflux/cmd/schemaflux@latest
12-pass compiler
Parse → slug → sort → enrich → taxonomy → relationships → graph → URL resolve → schema → validate → emit. Each pass is composable.
Multiple backends
HTML (with templates), JSON, RSS, XML sitemaps, search indexes, and custom backends via the Go API.
Template engine
Go templates with access to the full IR graph. Cross-document relationships, backlinks, and taxonomy pages generated automatically.
Eval dataset compilation
Compile annotated markdown into structured eval input datasets for matchspec.
Sub-second builds
~2,000 entities compiled to ~2,300 pages in under 500ms. Incremental builds coming.
Zero dependencies
Single static binary. Built on mist-go with no runtime requirements.
Example
schemaflux
$ schemaflux build --config schemaflux.yml
parsing 342 files
pass 1/12 slugify
pass 2/12 sort
pass 3/12 enrich
pass 4/12 taxonomy
pass 5/12 relationships
pass 6/12 graph
pass 7/12 url-resolve
pass 8/12 schema-gen
pass 9/12 validate
pass 10/12 emit-html
pass 11/12 emit-json
pass 12/12 emit-sitemap
output 412 files in _site/
built in 487ms
Most static site generators map files to pages. schemaflux compiles files to a graph first.
All relationships, taxonomy pages, cross-document backlinks, and search indexes are computed from the same IR before any output is written. Templates can traverse the full graph. The validation pass runs before emission — broken links and missing required fields fail the build.
The same pipeline that emits HTML also emits JSON, RSS, and sitemaps. One source, multiple outputs.