Lovkart · før du stoler på svaret

What Lovkart is — and where its edges are

Lovkart is a map of Norwegian primary law: the statutes, regulations, Supreme Court decisions, and the preparatory works behind them — plus the links between them. It's built to help you find and read the source. It is not legal advice, and knowing its limits is part of using it well.

01What's inside

Sources: Lovdata (lover & forskrifter), Høyesterett via domstol.no (avgjørelser), Stortinget & Regjeringen (forarbeider). Figures verified live against the corpus, 2026-08-09.

What makes it a map rather than a search box is the connections: 15,693 links from decisions to the law sections they cite, 17,655 to the preparatory works, and 14,845 amendment relationships between acts. You can walk from a statute § to the cases that cite it, to the reform that introduced it, to the regulation that implements it.

02A research tool, not legal advice

Lovkart helps you locate primary sources and quote them accurately. It does not give legal advice, predict outcomes, or tell you what the law means for your situation. It is not a substitute for Lovdata Pro or for a qualified lawyer. Treat every answer as a pointer to a source you then read and verify yourself.

03What it can't do

• Supreme Court only. No lower courts (lagmannsrett), no tribunals (nemnder). Høyesterett publishes ~240 of 2,000+ matters a year — the full picture is Lovdata Pro.
• No "still good law" check. Citing ≠ overruled or confirmed. Treatment analysis is out of scope.
• Historical law text is partial. Archived versions from 2014 onward for ~671 laws (via as_of); older and authoritative consolidations are Lovdata Pro.
• No commentary or foreign law. No lovkommentarer / juridisk litteratur (Juridika, Karnov); no EØS/EU or EMD sources; no local (lokale) regulations.
• Older case citations may not resolve. Pre-2016 "Rt. YYYY s. NNNN" references usually don't map directly — search by subject or party instead; the decision is often in the corpus under its HR-id.
• Some decisions are summary-only or stubs. Every decision is labelled with its text coverage. Never state what a case held from a summary or a stub — read the full text.
• Search finds words, not meaning. Full-text search is keyword discovery, not semantic search — matching words are a lead, not authority.

04How to use it responsibly

1. Follow the link.Every source comes with its lovdata.no / domstol.no / stortinget.no link. Open it for anything you'll act on.
2. Don't take a holding on faith.The tool tells you a decision cites a provision — not what it decided. Read the decision text before you state what it held.
3. A snippet isn't the source.Search fragments are for discovery. Fetch the full provision or decision and quote from that.
4. Use a capable model.The tools are validated with Claude Sonnet-class models and above; weaker models tend to drop the sourcing discipline.
5. When it matters, verify and ask a lawyer.For anything with legal consequence, confirm at the source and get professional advice.

05Your data & what's served

This private test runs in public-sources mode, and your data footprint is deliberately small:

• Public legal sources — served in full. The verbatim text of laws, regulations, and preparatory works, each with its lovdata.no / stortinget.no link.
• Court decisions — metadata and citation graph only. Case number, date, title, and the links to and from other sources — plus the domstol.no link to read the decision itself. The full text, summaries, judges, and legal area are withheld.
• Your queries — not logged. We don't store the content of your questions; standard server logs record only request metadata (IP, timestamp).