Croatian law, in a form software can actually use

Statutes, consolidated law and court decisions are published across three separate portals, in formats built for reading rather than querying. ProIure ingests all three continuously and serves them as one searchable corpus, with the statutory date arithmetic that legal work actually turns on. It is the data layer under lawyer.ai.

The law is public. That is not the same as being usable.

In production, refreshed continuously
500K+
legal documents

Gazette statutes, consolidated laws and court decisions from three separate portals, served as one searchable corpus.

In production

Running continuously, refreshed by scheduled ingest jobs. Available as a licensed API and as an MCP server.

The problem

Three portals, three formats, and none of them answer a question.

Croatian legal sources are public, which is not the same as usable. The official gazette publishes statutes as issues, not as law you can query. Consolidated texts live somewhere else again, and case law sits in a separate search portal. A lawyer checking whether a rule still applies, and what the courts have done with it, is reading three systems and reconciling them by hand.

  • Statutes arrive as gazette issues, so "what does this article say today" is a reading exercise, not a lookup.
  • Consolidated texts are convenient but unofficial, so authoritative work still has to go back to the gazette.
  • Case law is searchable only one decision at a time, with no way to ask what cites what.

What it does

One corpus, kept current, with the maths that matters.

ProIure ingests all three sources on a schedule and normalises them into a single full-text-searchable corpus with document history, so you can ask what changed and when. On top of the corpus sit pure-compute tools: statutory deadline calculation under the Civil Procedure Act, which has to respect non-working days, and citation formatting and checking, which catches references to provisions that no longer exist.

  • Over half a million legal documents in one place: gazette statutes, consolidated laws and court decisions.
  • Full-text search across the whole corpus, plus document history and change comparison.
  • Statutory deadline calculation that accounts for non-working days, rather than counting calendar days.
  • Citation formatting and checking against the live corpus, so a dead reference is caught before filing.
  • Watchlists on specific regulations, so a change to a rule you rely on surfaces on its own.

Why it exists

An assistant that cannot cite is worse than no assistant.

A general-purpose model asked about Croatian law will produce something confident and frequently wrong, because the law it needs was never reliably in its training data and it has no way to check itself. ProIure exists so the answer comes from the actual published text, with the source named. It is served as an MCP server as well as an API, so Claude and other assistants read the real corpus rather than recalling an approximation of it.

  • 19 MCP tools: regulation lookup and history, court-decision search, and the legal compute tools.
  • Every result carries its source, because the source portals require attribution on reuse.
  • Authoritative text traces to the official gazette; consolidated texts are labelled as the convenience they are.

The detail, in one place.

Corpus
Over half a million documents: gazette statutes, consolidated laws and court decisions
Sources
Narodne novine, Zakon.hr, odluke.sudovi.hr
Refresh
Scheduled ingest jobs, running continuously
Search
Croatian-language full-text search across the whole corpus
Legal compute
Statutory deadline calculation, citation formatting and checking
Interfaces
Licensed REST API and an MCP server with 19 tools
Platform
.NET 10, Azure SQL with Croatian full-text search, Azure Blob storage

Sources

ProIure aggregates public Croatian legal sources and names them on every result, because their terms require it and because a legal answer without a source is not an answer.

  • Narodne novine — the official gazette, and the authoritative text
  • Zakon.hr — consolidated texts, a convenience and not authoritative
  • odluke.sudovi.hr — the court-decisions portal

Frequently asked questions

Which jurisdiction does ProIure cover?
Croatia. It ingests the Croatian official gazette (Narodne novine), consolidated legislation from Zakon.hr, and the court-decisions portal odluke.sudovi.hr. The ingestion architecture is not jurisdiction-specific, but Croatian law is what is live today.
How is this different from searching the public portals?
The portals are three separate systems, each searchable one document at a time, and none of them computes anything. ProIure puts all three in one corpus with full-text search and document history, and adds statutory deadline calculation and citation checking on top. It also exposes the whole thing to software, which the portals do not.
What is the relationship between ProIure and lawyer.ai?
ProIure is the data layer; lawyer.ai is the practice-management system a firm actually works in. lawyer.ai reads the ProIure corpus read-only. If you run a law firm you almost certainly want lawyer.ai. If you are building something of your own and need the legal data underneath it, you want ProIure.
Can an AI assistant use ProIure directly?
Yes. It ships as an MCP server with 19 tools, so Claude and other MCP-capable assistants can search regulations, look up consolidated text, search court decisions, calculate statutory deadlines and check citations. Every result carries its source so the assistant can attribute what it tells you.
Is the consolidated text authoritative?
No, and the service is explicit about that. Consolidated texts are a convenience for reading. For authoritative wording, the answer traces to the official publication in Narodne novine, and ProIure returns both so the distinction is never lost.