Steuerberaterverzeichnis Scraper — German Tax Adviser Leads with Email & Phone

Export German tax advisers (Steuerberater), Steuerbevollmächtigte and tax-advisory firms (Steuerberatungsgesellschaften) from the official Steuerberaterverzeichnis kept by the Bundessteuerberaterkammer (BStBK) — with the business email, phone, mobile, website and office address each member publishes there. Search by postcode, city, name, regional chamber, or the whole of Germany.

• ✉️ Email on 97.5% of records, phone or mobile on 82%, website on 28% (measured on 17,882 live records — see the per-chamber table below)
• 🏛️ Official source: the Bundessteuerberaterkammer's register of every German tax adviser
• 🗺️ Complete coverage: the register refuses broad searches ("Zu viele mögliche Treffer"); this actor narrows every refused search digit by digit and checks its row count against the register's own count for every cell
• 🔑 Safe ID (beSt electronic mailbox id) on every record, plus appointment date, chamber, firm management and partners
• 💶 $4 per 1,000 records, charged only for records delivered — no start fee, no charge for filtered-out or duplicate records

Who uses this

• Tax & accounting software vendors (DATEV alternatives, bookkeeping, payroll, e-invoicing / XRechnung tools) building a partner or sales pipeline of every Kanzlei in a region
• Legal-tech, fintech and HR/payroll partner programmes recruiting Steuerberater as channel partners
• Agencies feeding Kanzlei addresses into enrichment and CRM workflows
• Compliance teams verifying that an adviser is registered (and with which chamber) before relying on them

What one record looks like

Real record from a verified run (Hamburg, 2026-09-23):

{
  "registerId": "44-DB-BC-92-53-1E-60-3A-30-48-4D-AF-87-B1-C3-71",
  "entityType": "COMPANY",
  "name": "Jacobsen + Confurius Rechtsanwälte Steuerberater Wirtschaftsprüfer Partnerschaftsgesellschaft mbB",
  "profession": null,
  "academicTitles": null,
  "legalForm": "PartG mbB",
  "addressAddition": null,
  "street": "Pinnasberg 47",
  "postalCode": "20359",
  "city": "Hamburg",
  "address": "Pinnasberg 47, 20359 Hamburg",
  "email": "info@jacobsen-confurius.de",
  "phone": "040 3020030",
  "mobile": null,
  "fax": "040 337820",
  "website": "www.jacobsen-confurius.de",
  "websiteUrl": "http://www.jacobsen-confurius.de",
  "hasEmail": true,
  "hasPhone": true,
  "hasContact": true,
  "safeId": "DE.BStBK.8f5d8ca5-41b6-44ad-be2d-92e61f2d80fd.2709",
  "dateType": "Anerkennungsdatum",
  "date": "2022-12-06",
  "chamber": "Hamburg",
  "chamberAsPublished": "Hamburg",
  "chamberAddress": "Kurze Mühren 3, 20095 Hamburg",
  "otherOffices": ["Eisenzahnstr. 64, 10709 Berlin-Wilmersdorf"],
  "foreignSeat": null,
  "foreignPrincipalPlace": null,
  "managementBody": [
    { "name": "Jens Borchardt", "profession": "Rechtsanwalt" },
    { "name": "Manfred Confurius", "profession": null }
  ],
  "authorizedPartners": [ { "name": "Manfred Confurius", "profession": null } ],
  "partners": [],
  "practiceRepresentatives": [],
  "detailUrl": "https://steuerberaterverzeichnis.berufs-org.de/details/44-DB-BC-92-53-1E-60-3A-30-48-4D-AF-87-B1-C3-71/?lang=de",
  "scrapedAt": "2026-09-23T04:24:55.119Z"
}

(The management and partner lists are shortened here; the real record lists all 15 and 11 names.) An individual adviser looks the same with entityType: "PERSON", a profession such as Steuerberater / Steuerberaterin / Steuerbevollmächtigter, academicTitles such as Dr., and — for employed advisers — the firm they work at in addressAddition.

Every field

Field	What it is
registerId	The register's own record id. Stable across runs; the dedupe key — no record is delivered or charged twice in a run
entityType	PERSON (Steuerberater/in, Steuerbevollmächtigte/r) or COMPANY (Berufsausübungsgesellschaft)
name	Full name of the adviser, or the firm name
profession	Berufsbezeichnung, e.g. Steuerberater, Steuerberaterin, Steuerbevollmächtigter (people only)
academicTitles	e.g. Dr., Dipl.-Kaufmann, as printed
legalForm	Firms only: GmbH, PartG mbB, GbR, …
addressAddition	Lines above the street — usually the firm an employed adviser works at
street, postalCode, city, address	Office (berufliche Niederlassung) address
email	Business email
phone, mobile, fax	Business numbers, exactly as printed
website, websiteUrl	Website as the member wrote it, and the register's own clickable form of it
hasEmail, hasPhone, hasContact	Booleans for fast filtering in a spreadsheet
safeId	Safe ID of the besonderes elektronisches Steuerberaterpostfach (beSt)
dateType, date	Which date the register gives (Bestelldatum = appointment, Anerkennungsdatum = firm recognition, Registrierungsdatum) and the date itself, ISO
chamber, chamberAsPublished, chamberAddress	The responsible Steuerberaterkammer (one of 21)
otherOffices	A firm's further advisory offices (Weitere Beratungsstellen)
foreignSeat, foreignPrincipalPlace	Foreign firms only: country of seat and town of principal establishment
managementBody, authorizedPartners, partners	Firms: named directors, representing partners and shareholders, each with profession where printed
practiceRepresentatives	An adviser's appointed Praxisvertreter: name and address
detailUrl	The record in the official register
scrapedAt	When this run read the record

Contact fill — measured, per chamber

Measured on 17,882 distinct records from five live runs on 2026-09-23 (postcode areas 0xxxx, 1xxxx, 20xxx, 80xxx, 657xx, 48143, 39104). "Phone" = landline; "phone or mobile" = either.

Chamber	Records	Email	Phone	Phone or mobile	Website	Any contact
Berlin	4,557	97.5%	84.7%	94.0%	33.8%	99.0%
München	3,415	96.4%	82.2%	94.9%	23.7%	98.7%
Sachsen	3,141	97.0%	59.2%	63.4%	37.0%	98.0%
Hamburg	2,658	99.7%	48.8%	57.7%	12.0%	99.9%
Brandenburg	1,478	99.7%	74.6%	91.9%	28.5%	99.9%
Mecklenburg-Vorpommern	934	94.5%	89.1%	96.6%	22.9%	98.4%
Sachsen-Anhalt	678	97.8%	90.7%	95.0%	47.5%	99.1%
Hessen	471	97.2%	50.7%	58.4%	23.1%	98.3%
Thüringen	376	99.5%	92.6%	94.4%	28.2%	99.5%
Westfalen-Lippe	117	86.3%	53.0%	61.5%	7.7%	89.7%
All 17,882		97.5%	73.0%	82.1%	28.1%	98.8%

Chambers not in the table (Düsseldorf, Köln, Niedersachsen, Nordbaden, Nürnberg, Rheinland-Pfalz, Saarland, Schleswig-Holstein, Stuttgart, Südbaden, Bremen) had fewer than 100 records in these runs; a second, smaller frame of 758 records across seven chambers (incl. Bremen 150, Westfalen-Lippe 116) read before the build showed the same shape: email 86–100%, phone 39–95%, website 8–41% per chamber. How often the phone field is filled is a property of the chamber, not of this actor.

Each chamber feeds the register automatically. A missing phone or website means the adviser did not give one to their chamber — not a parsing gap: on every run, RUN_SUMMARY.labelPresentToValueEmitted counts how often the register printed each label against how often a value was emitted.

Input

Input	What it does
postcodes	1–5 digit postcodes or prefixes. 80331 = one postcode; 803 = all of 803xx; 8 = all of 8xxxx
cities	Town names (Münster, Frankfurt am Main) — the register matches the start of the name
names	Surnames or firm names, * / ? wildcards allowed (Mül*)
allGermany	The whole register (109,109 records on 2026-09-23). Ignores the three lists above. Paced by the register's quota — see Big jobs below
chamber	Keep only one of the 21 Steuerberaterkammern. Applied on the register's side where possible, so a region with no member of that chamber costs no record requests
entityType	ALL, PERSON (individual advisers) or COMPANY (firms). Decided from the result list, so it costs nothing extra
requireEmail / requireContact	Deliver (and charge for) only records with an email / with any contact
maxResults	Stop after this many delivered records (default 80; 0 = no cap)
detailConcurrency	Parallel requests (default 8). The actor halves it automatically when the register answers HTTP 429
residentialFallback	If the register blocks this run for a minute or more (its limit is shared with other jobs), continue on Apify RESIDENTIAL proxy instead of waiting. Default on; no effect on price
proxyConfiguration	Off by default: the register has no anti-bot and answers Apify's servers directly

Leave everything empty to run the small demo scope (postcode 65719, Hofheim am Taunus).

Examples

{ "postcodes": ["80331", "80333", "80335"], "maxResults": 0 }

Every tax adviser in central Munich.

{ "allGermany": true, "chamber": "Westfalen-Lippe", "requireEmail": true, "maxResults": 0 }

Every member of the Steuerberaterkammer Westfalen-Lippe that publishes an email.

{ "cities": ["Leipzig"], "entityType": "COMPANY", "maxResults": 500 }

Tax-advisory firms in Leipzig, with their directors and partners.

Pricing

$0.004 per record delivered ($4 per 1,000). One event, advisor-record, charged only when a record lands in your dataset. Not charged: records removed by your chamber / entityType / requireEmail / requireContact filters, a record already delivered in the same run, searches that match nobody, or anything that could not be fetched. No start fee.

Job	Records	Cost
Demo scope (empty input)	80	$0.32
All of Hamburg 20xxx	2,656	$10.62
Munich 80xxx + Berlin 10xxx	6,135	$24.54
The whole register	109,109	$436.44

Set Max results or Apify's maximum cost per run to cap a job; the actor stops cleanly at either and never delivers a record it could not charge for.

How it works (and why the count is right)

1. The register answers a search with every matching record on one page ("Treffer: 671 von 671") — or refuses it outright as too broad. A refused search is narrowed along the postcode: 8?? → 80* → 803 → 8033 → 80331. A full 5-digit postcode has always been answered (largest seen: 10117 Berlin, 671 records).
2. For every answered cell the actor floor-asserts the rows it parsed against the register's own count. The register's ordering is not deterministic on big cells — the same request for 20354 returned 660, 659 and 660 distinct records under an unchanged "661 von 661" — so a short cell is asked again and the answers are unioned by register id until they reach the register's count. Cells that never reach it are listed in RUN_SUMMARY.
3. Each record's page is opened for the contact data (the result list carries only name and address), deduplicated by registerId before charging.
4. RUN_SUMMARY keeps the kinds of nothing apart: cells the register answered with no records, searches it refused as invalid, cells it refused as too broad (and how they were narrowed), cells whose count and rows disagreed, and records that stayed unreachable. None of them are charged.

Verified completeness: an independent census walk of every postcode cell on 2026-09-23 counted 109,109 register rows (109,103 distinct ids — the six-record gap is the non-deterministic ordering above). The live actor matched the register's count exactly on every area checked: 80xxx + 10xxx 6,135 = 6,135, 1xxxx 6,802 = 6,802.

Big jobs and the register's rate limit

The register answers 429 Zu viele Anfragen with a Retry-After once a client group has sent roughly 8,900 requests in a short window, and knocking during that block makes it longer. The quota is shared: ten parallel runs on ten different Apify IPs were blocked in the same second, so running several jobs at once is not faster than one. The actor therefore spends a burst of 6,000 requests at full speed — a city, a postcode area or a whole chamber's region of up to ~5,500 records finishes inside it, in minutes — and then paces itself at 5 requests per second, honouring any Retry-After in full and pausing every request while it lasts. A record never costs more because of this; a big job just takes longer: the whole register (≈116,000 requests) takes about 6–7 hours in one run (the default run timeout is 8 hours). If a search area still cannot be reached, the run ends FAILED with the areas named rather than reporting a partial export as complete — and the records it did deliver are real and were charged.

Honest limits

• Contact is only what the adviser gave their chamber. About 4% of records have no email; websites are published by roughly a quarter. The actor never guesses a domain or completes a number.
• Malformed entries are not repaired. The register prints a few emails like x@firm,de, x@me-com or x@firm .de; those are not emitted (they are listed in RUN_SUMMARY.emailsRejectedAsMalformed). A pasted E-Mail: label or a trailing full stop around a valid address is trimmed; umlaut domains (…@steuerkanzlei-hübner.de) are real and are kept.
• Sachsen-Anhalt is printed by the register only as "Körperschaft des öffentlichen Rechts". chamber resolves it to Sachsen-Anhalt from the chamber's own Magdeburg address; chamberAsPublished keeps the printed text.
• Professional bans (Berufs-/Vertretungsverbote) are part of the register entry, but none was printed on any record we read, so there is no field for them yet. If the register prints one, it appears in RUN_SUMMARY.rareNoticeBlocks rather than being dropped.
• People in firm lists (managementBody, partners) carry a profession only where the register prints one.
• The register rate-limits (see Big jobs). The actor backs off run-wide and never keeps knocking during a block; a whole-Germany export therefore takes hours, and runs started in parallel share the same quota.

Source

The Steuerberaterverzeichnis of the Bundessteuerberaterkammer, https://steuerberaterverzeichnis.berufs-org.de/.

FAQ

Can I get every tax adviser in Germany? Yes — allGermany: true, maxResults: 0, in ONE run (about 6–7 hours, paced by the register's quota; parallel runs share that quota and are not faster). 512 MB is enough (peak measured: 296 MB on 6,135 records).

Why not a Google Maps or Gelbe Seiten scrape? Those list whoever bought a listing. This is the official register: every appointed Steuerberater is in it, with the email they registered with their chamber, and nobody else is.

How fresh is it? Each run reads the live register; scrapedAt is on every record.

Something looks wrong? Open an issue with the run id. RUN_SUMMARY holds what the actor saw.