There is no official, sales-friendly API for the German, Austrian or Swiss registers. There is, however, a great deal of public, authoritative data if you know where each fact lives, and the discipline to say not established when it does not.
The problem every DACH seller has met
Sales intelligence built for the US assumes a single register with a public API and a culture of companies publishing everything. DACH works differently. Germany's commercial register is organised by local court, Austria's Firmenbuch is a paid extract service, and Switzerland keeps cantonal registers behind a federal index. None of them offers an API designed for a sales team to query by domain, and scraping them is neither reliable nor, in some cases, permitted.
The temptation is to fall back on a broker's database, which is stale, unsourced and priced per record. The better answer is to understand what is freely and legitimately available, and to build research around it.
What is freely available
Germany: the common register portal allows anyone to retrieve register entries and current company data free of charge, and companies are obliged to publish a legal notice on their website with the register court, HRB number and VAT ID. Financial statements are published in the company register and the Bundesanzeiger.
Austria: the Firmenbuch is searchable, but extracts are generally chargeable; the legal-notice obligation under the ECG and UGB still gives you the company's legal name, seat and Firmenbuch number from its own website. Switzerland: the federal Zefix index is free to search and exposes the UID and the cantonal register entry, and cantonal registers publish daily notices of changes. All three jurisdictions publish VAT registration through the EU VIES system or, for Switzerland, the UID register.
What the legal notice, GLEIF and VIES give you
The legal notice gives you identity in the register's own vocabulary: legal name, legal form, seat, court and number, VAT ID, and usually the managing directors. It is the bridge from a domain to an entity, and Caicle reads it in the first research wave.
GLEIF, the global legal-entity register, gives you the LEI and the ownership relationships between LEI-holding entities: direct parent, ultimate parent, children. For any company in a group, this is where the question of who actually signs gets answered. VIES gives you one thing, reliably: whether a VAT ID is valid and, for many member states, the registered name and address behind it. That is a confirmation, not a source of new facts, and Caicle treats it as one.
A confirmed VAT ID, a register number from the legal notice and an LEI from GLEIF are three independent sources agreeing on one entity. That agreement is what a confidence score is made of.
Where Caicle says not established, on purpose
Some facts are simply not public in DACH, or not public in a form that can be established without a paid extract. Exact headcount for a GmbH below the disclosure thresholds is one. Revenue for most private companies is another; where a public source states a figure, the dossier shows it as an estimate with the source, and otherwise the field stays empty. Shareholder lists are filed but not machine-readable in a way Caicle will vouch for.
The same goes for people. Executives in professional roles come from registers and legal notices. Private-life data is out of scope by system rule, and a person who cannot be established from a professional source is not listed.
Precedence, and why the blanks are the point
When sources disagree, precedence decides: a register value at 0.95 beats the legal notice at 0.90, which beats Wikidata at 0.70, which beats the company's own website copy at 0.60. The dossier shows every field's source, URL, confidence and the date it was observed, and a list of the fields no source could confirm sits alongside them.
That list is not a weakness of DACH data. It is the honest shape of it, and it is what lets a rep say, in a meeting with a German buyer, exactly which facts came from the register and which did not. A dossier that is right and incomplete is worth more than one that is complete and partly invented.