The corpus · Measured live

The shared code library, in numbers.

Stellify's reuse loop only works if there's a real corpus behind it. This page is that corpus, measured: how much production-ready Laravel code exists, what it covers, how much duplication referencing avoids, and which public modules other projects actually wire in. Every figure is computed from the live database and refreshed daily — nothing here is curated copy.

Why reuse beats regeneration is measured separately — the economics of code reuse.

Code units in the corpus
31,845
Duplicate statements avoided
1,547
Cross-project code references
260
Projects reusing shared code
27

Measured live from the corpus · refreshed daily

Composition

What a corpus of structured code is made of.

Code on Stellify is stored as structured data — files reference methods, methods reference statements, statements reference clauses — which is what makes statement-level reuse and deduplication possible.

files
653
methods
1,308
statements
10,449
clauses
19,435
Growth · Last 30 days

The corpus, day by day.

Code units in the corpus
Jul 18 Aug 16
Live reference edges
Jul 18 Aug 16
Most referenced

The code other projects wire in most.

Only units from public Constellation projects are named here; private projects are never listed, whatever their adoption. "Reused by" counts distinct projects holding a live reference or fork of the unit.

Unit Kind From Reused by
entries method Audit Log 10 projects
down method Auth & Teams 10 projects
prune method Audit Log 10 projects
up method Auth & Teams 10 projects
record method Audit Log 10 projects
register method Auth & Teams 7 projects
userFromToken method Auth & Teams 7 projects
render method Transactional Email 7 projects
issueToken method Auth & Teams 7 projects
buildPayload method Transactional Email 7 projects

How this page is measured: every figure is an aggregate computed from the live corpus and refreshed daily — the real, unrounded numbers, nothing padded or curated. A counter only appears once it has grown past a fixed minimum, so what you see here is earned. Private projects contribute to corpus-level totals but are never named, and no user code or search content is ever shown.