If you work with Swedish companies — sales prospecting, KYB checks, credit or market research — you've probably met allabolag.se, ratsit or proff. They're great websites, but scraping them is fragile and against their terms, and paid data providers charge per seat.
Since 2025 there's a better option: Bolagsverket, the Swedish Companies Registration Office, publishes its register and digitally filed annual reports free of charge under the EU High-Value Datasets regulation (2023/138). This article shows three ways to use it, from DIY to a one-call API.
1. What's actually available
| Source | What you get | Access |
|---|---|---|
| Bolagsverket bulk file | every registered organisation: name, legal form, status, registration date, postal address, business description | ~250 MB zip, weekly, no login |
| SCB bulk file | statistics-register companion (industry codes etc.) | ~70 MB zip, no login |
| Bolagsverket "Värdefulla datamängder" API | organisation lookup by org number, list + download of annual reports (iXBRL) | free, OAuth client credentials, 60 req/min |
| ESEF filings (filings.xbrl.org) | annual reports of listed companies as xBRL-JSON | open |
| GLEIF LEI register | legal name, address, legal form for companies with an LEI | CC0 API |
What's not in the open data: board members, owners, phone numbers and emails.
2. DIY: load the whole register with pandas
import pandas as pd df = pd.read_csv( "https://vardefulla-datamangder.bolagsverket.se/bolagsverket/bolagsverket_bulkfil.zip", sep=";", encoding="utf-8", quotechar='"', escapechar="\\", compression="zip", low_memory=False, on_bad_lines="warn") (Credit: this loading recipe is from Pierre Mesure's oppna-bolagsdata repo.)
Things to know: multi-valued fields are packed as value$type|value$type; the address is street$c/o$city$postcode$country; and the server can drop long downloads, so resume with an HTTP Range header if you script it.
3. DIY: parse revenue and profit from an annual report
Digitally filed Swedish annual reports are inline XBRL: an XHTML page where every number is tagged. For small companies (K2/K3) the tags come from the Swedish taxonomy:
| Concept | Meaning |
|---|---|
se-gen-base:Nettoomsattning | revenue |
se-gen-base:Rorelseresultat | operating profit |
se-gen-base:ResultatEfterFinansiellaPoster | profit after financial items |
se-gen-base:AretsResultat | net profit |
se-gen-base:EgetKapital | equity |
se-gen-base:Tillgangar | total assets |
se-gen-base:MedelantaletAnstallda | average employees |
Listed companies file ESEF reports with ifrs-full concepts instead (Revenue, ProfitLossFromOperatingActivities, ProfitLoss, Equity, Assets). Watch the scale and sign attributes on each ix:nonFraction tag, and the number format (numcommadot vs numdotcomma).
Minimal extraction with lxml:
from lxml import etree IX = "http://www.xbrl.org/2013/inlineXBRL" def facts(xhtml_bytes): root = etree.fromstring(xhtml_bytes) for el in root.iter(f"{{{IX}}}nonFraction"): txt = "".join(el.itertext()).replace(" ", "").replace("\u00a0", "").replace(",", ".") if not txt: continue val = float(txt) * 10 ** int(el.get("scale", "0")) if el.get("sign") == "-": val = -val yield el.get("name"), el.get("contextRef"), val You then pick the context that matches the latest fiscal year (contexts are defined in xbrli:context elements).
4. The one-call option
If you'd rather not maintain parsers, download jobs and rate limiters, I packaged all of the above as an Apify actor (disclosure: I built it): Swedish Company Data & Financials.
from apify_client import ApifyClient client = ApifyClient("<APIFY_TOKEN>") run = client.actor("stimulating_stilbite/swedish-company-data-financials").call( run_input={"orgNumbers": ["556016-0680", "556042-7220"], "includeFinancials": True}) for c in client.dataset(run["defaultDatasetId"]).iterate_items(): f = c["financials"] or {} print(c["name"], f.get("fiscalYearEnd"), f.get("revenue"), f.get("netProfit")) Real output from 18 Nov 2026:
Telefonaktiebolaget LM Ericsson 2025-12-31 236681000000 28714000000 H & M Hennes & Mauritz AB 2025-11-30 228285000000 12085000000 It also works from no-code tools (Make, Zapier, n8n, Google Sheets) and from AI agents via the Apify MCP server. Current limits, honestly: financials cover ESEF filers (listed issuers) today, with small-company K2/K3 parsing being switched on; registry data for every organisation is available via the bulk-file option.
Wrap-up
Swedish company data used to be a paid product. Now it's public infrastructure — the only work left is parsing. Use the raw files if you have the time, or the actor if you don't. Questions or field requests: comment below.