Open Laws Foundation

An interoperability and provenance layer for the world’s legislation.

Most legal-data projects scrape statutes and dump them into a convenient format. That throws away the structure: cross-references, temporal validity, the relationships between acts. And that structure is the part that actually makes legislation useful to machines. We do the opposite.

Read the spec GitHub ↗︎

Stable identifier olf:it/legge/2019/123/art_3

What we are

A profile of Akoma Ntoso (the OASIS legal-document standard), not a new format. We speak the language the EU, the UN, and national gazettes already speak.
A set of per-jurisdiction adapters that ingest official sources and emit validated Akoma Ntoso.
A semantic, type-aware diff for legislation: it tells you what kind of change happened. Text amended, entry into force changed, cosmetic fix.
A conformance suite that defines what “correct” means, so independent contributors stay interoperable without ever talking to each other.

What we are not

Not another scraped Markdown/PDF corpus.
Not a replacement for Akoma Ntoso, ELI, or any national system.
Not (yet) a legal entity. This is an open-source project. We do not solicit or accept donations on behalf of a “foundation” that has not been incorporated.

The one-sentence design rule

Normalize the metadata: identity, time, citations. Never the content. If two lawyers from different countries would argue about how to model it, it stays native. If they agree it exists in both systems, we normalize it.

Entry into force? Exists everywhere → normalize. comma vs section structure? They’d argue → native.

The repositories

Four pieces. Start with the spec; everything else conforms to it.

spec Apache-2.0

AKN4OLF profile + conformance suite

The profile of Akoma Ntoso and the tests every adapter must pass. The entry point. Read this first.

pipeline Apache-2.0

Adapters & ingest workflow

Per-jurisdiction adapters and the ingest → validate → publish workflow that turns official sources into validated Akoma Ntoso.

diff Apache-2.0

Semantic, type-aware differ

The semantic differ and changeset model. It understands Akoma Ntoso structure, so it knows the difference between an amendment and a typo fix.

archive CC0-1.0

Canonical corpus

The canonical body of legislation, in Akoma Ntoso. Generated by the pipeline, not hand-edited. Public domain.

Status

Early. We are building the first two jurisdictions, Italy and France, to prove that one model holds across two legal traditions. Civil-law structure, two different gazettes, one identifier scheme.

Maintain legal data for a jurisdiction and want it represented here? Open an issue in pipeline ↗︎

it Italy in progress
fr France in progress
·· Your jurisdiction open