Telefonbuch.de Scraper — German Names, Addresses & Phones

Scrape Germany's phone directory: name, street, postcode, city — and the full phone number, reconstructed from markup that is deliberately built to prevent exactly that.

HTTP-only, no account, no browser.

Input

Field	Type	Default	Description
queries	array	required	name, city (Schmidt, Berlin), name/city, or a pasted search URL
maxResultsPerQuery	integer	100	Entries come 10 to a page
maxPages	integer	10	Stops early at the end of results
maxRounds	integer	2	TLS ladder repeats on a refused page
proxyConfiguration	object	RESIDENTIAL DE	German directory — use a German exit

The phone number is hidden, twice

This is what the actor is actually for. The site splits and salts its phone numbers so that the obvious extraction returns something wrong, and it does it two different ways on the same results page:

1. Truncated itemprop.

<span itemprop="telephone">0162 2 13 2<span style="display:none">…</span></span>1 39

Reading the itemprop element gives 0162 2 13 2… — 7 of 11 digits, not dialable. The rest of the number sits in the text node after the span closes. The detail page truncates it the same way, so following the link does not help.

2. Decoy digits in hidden spans. Other entries carry no itemprop at all:

0<span class="hide"> 3</span>3<span class="hide"> 96</span>0…

Read naively that yields a 26-digit number that is simply wrong — and, unlike a truncated one, it looks complete. It passes any "is the field populated?" check.

One treatment defeats both: drop every hidden span, then read the container. Each number is then cross-checked against the site's own phoneTo markup and the comparison is reported — 87 of 87 matched in testing, and every row carries phoneFromComment so you can audit it yourself.

Output

{
  "_input": "Schmidt, Berlin",
  "_source": "S1-vcard",
  "_scrapedAt": "2026-09-07T11:04:06Z",
  "_page": 1,
  "_phoneSource": "dom",              // or "comment" if the fallback was used

  "name": "Algert-Schmidt Barbara",
  "phone": "0162 2 13 21 39",
  "phones": ["0162 2 13 21 39"],      // businesses often list several
  "phoneDigits": "01622132139",
  "phoneFromComment": "0162 2 13 21 39",
  "streetAddress": "Luckenwalder Str. 78",
  "postalCode": "12629",
  "city": "Berlin",
  "searchTerm": "Schmidt", "searchLocation": "Berlin",
  "detailUrl": "https://kontakt-1.dastelefonbuch.de/…"
}

In a 90-entry run: phone on 90/90, name 89/90, detail URL 90/90.

Known limits

Limit	Detail
Addresses are partial	Street on ~74/90, postcode and city on ~80/90 — some entries genuinely publish a number without an address
German exit needed	Reachability was measured 4/4 from a DE residential exit; other geos were not tested
No email	The directory does not publish email addresses
Watch _phoneSource	If it starts reporting comment or none in bulk, the reconstruction has drifted and the run log warns about blank phones