Brutto
Launch app
Norwegian company intelligence

Put the Norwegian market in your agent's hands

Registry records, accounts, ownership, grants, procurement, filings and web evidence — joined into one company dataset. Reach it three ways: the web portal, MCP tools, or a direct download.

Join waitlistSee the surfaces
Built on official Norwegian registry data and public sources.
founder-owned pump service companies with recurring revenue
CompanyRev mGrowthFit
Nordic Pumps AS428+22%0.96
Vest Servicepartner AS184+9%0.92
Arctic Flow Control AS96+31%0.88
Fjord Hydraulikk AS61+6%0.84
recurring revenuepumpeservicefounder-owned147 matches
One dataset, every source normalized
BRREGREGNSKAPAKSJONÆRERDOFFINENOVAPATENTSTYRETWEB
Three ways in

The same data, wherever the work happens

Brutto is one joined dataset with three front doors. People use the portal; agents call MCP; engineers take the whole thing as a download. Nothing is stitched back together by hand.

Web portal

for analysts

Screen all 1.2M companies, draw a market on the map, open a company profile with financials, ownership and cited evidence — no SQL.

founder-owned pump service companies with recurring revenue
CompanyRev mGrowthFit
Nordic Pumps AS428+22%0.96
Vest Servicepartner AS184+9%0.92
Arctic Flow Control AS96+31%0.88
Fjord Hydraulikk AS61+6%0.84
recurring revenuepumpeservicefounder-owned147 matches

MCP

for agents

Seven deterministic tools your agent calls directly — resolve facets, screen, find comparables, build market maps and verify every claim against source evidence.

brutto-intel · MCP
› search_companies({
    plan: "pump service · recurring rev",
    min_revenue: 25_000_000
})

  ↳ 147 companies, scored + cited
› get_evidence(orgnr)  → source snippet + url
resolve_facetsfind_comparablesmarket_mapget_company_profile

Direct download

for your stack

Take the whole dataset — Postgres dump, Parquet or DuckDB — plus an agent-native bundle that runs the same tools against your own copy.

Dataset bundlerelease 2026.06
brutto_full.pgdump
Postgres 18 · pgvector · all layers
tables/*.parquet
L1–L4 + cited evidence
brutto.duckdb
Offline, laptop-ready
screen.xlsx
Any shortlist, one click
Ships with CLAUDE.md, a local MCP server and the brutto CLI.
Jobs to be done

Point an agent at a market. Get a defensible shortlist.

Every step below is a tool call, not a prompt guess. The agent screens, ranks, drills to evidence and hands you a list you can put in front of an investment committee.

Start automating research
  1. 1Resolve a thesis into structured facets and semantic terms
  2. 2Screen every company in the segment, scored
  3. 3Filter on ownership, size, growth and geography
  4. 4Pull profiles — financials, owners, roles
  5. 5Attach a source snippet to every claim
  6. 6Hand the shortlist back as a file or a live answer
What's inside

From source row to shortlist, joined

Screens, profiles, exports, comparables and market maps all run off the same dataset, with the underlying source rows always one click away.

Company directory

Every Norwegian company, group, subunit and operating entity in one resolvable graph.

Financials

15+ years of accounts, ratios, growth and statement lines with year-correct currency.

Ownership & roles

Shareholders back to 2009, board roles, ultimate parents and group structures.

Filings & evidence

Annual accounts, grants, procurement, IP rights and cleaned web snippets — each cited.

Market maps

Exhaustive segments, comparables, activity shares and concentration per market.

Exports

Any screen or shortlist to Excel and CSV, ready for models, memos and trackers.

Built for

Different desks. One source of truth.

Sourcing depth for

Private equity

Find proprietary platforms and add-ons by niche activity, size, geography, ownership and growth.

Better lists for

M&A advisors

Build buyer lists, comparable sets and market maps without rebuilding the same workbook each time.

Evidence-first work for

Consultants

Size markets, map value chains and verify every claim from source-linked company facts.

By the numbers

The math behind the product

1.2M
Norwegian companies in the screen
Brønnøysundregistrene + source tables
15+
Years of filed accounts for historical analysis
Regnskapsregisteret + parsed reports
2009–25
Shareholder history wired into profiles
Skatteetaten shareholder books
7
Deterministic MCP tools for agents
brutto-intel MCP server
“Every thesis used to start with a week of gathering accounts, ownership and public records. Now the source work is instant — and MCP means our agents use the exact facts we trust.”
Illustrative customer
Corporate development lead
Learn more

Frequently asked questions

What is Brutto?

A company-intelligence dataset for Norway. It joins public company records, accounts, ownership, roles, grants, procurement, filings, IP and web evidence into one searchable product — reachable in the portal, over MCP, or as a download.

How is it different from public portals?

Public portals expose one source at a time. Brutto connects the sources into company profiles, market maps, comparables and exports, so you can ask one question across the whole market — and an agent can too.

Can we download the whole dataset?

Yes. Brutto ships as a Postgres dump, Parquet, or a DuckDB file that works offline on a laptop, alongside an agent-native bundle with a local MCP server and CLI.

Why does MCP matter?

MCP lets an AI tool call deterministic Brutto tools for screens, profiles, comparables, market maps and evidence. The agent reasons over real results instead of guessing the underlying facts.

Is the data auditable?

Every claim carries a source row, filing or cited evidence snippet, so you can verify a fact before it goes into a model, memo or deck.

Hand the source work to your agents.

One dataset for the whole Norwegian market — in the portal, over MCP, or downloaded into your stack. You keep the judgement calls.

Join waitlist