New York State Bar Attorney Scraper — License Lookup & Leads

Every attorney on the New York bar register, from the state's own open-data feed at data.ny.gov. One row per attorney registration: registration (bar) number, full name split into first/middle/last, firm, firm head-count, phone, business address, city, state, ZIP, county, law school, year admitted, years licensed, judicial department, registration status and next registration due. Filter by status, county, city, address state, admission-year cohort or firm size — or paste registration numbers and surnames to verify specific attorneys. No login, no API key, no CAPTCHA, no proxy needed.

432,690 registrations. 293,648 currently registered. 251,320 of those carry a phone or a street address — every count measured live on 2026-08-12, not quoted from a marketing page. Phone fill on that default slice is 100%.

🚀 Quick start

The Actor ships prefilled with a working demo — statuses: Currently registered · county: New York (Manhattan) · 100 rows. Press Start with zero edits and you get 100 currently-registered Manhattan attorneys with firm, firm head-count, phone, street address, ZIP, county, law school and admission year. There are 77,761 attorneys behind that filter; you get the first 100 of them — raise maxItems when you want the rest.

{
  "statuses": ["Currently registered"],
  "counties": ["New York"],
  "maxItems": 100
}

Three more inputs worth copying:

// 1. The legal-SaaS ICP: solo and small-firm attorneys, statewide, reachable
{ "statuses": ["Currently registered"], "withContactOnly": true, "maxFirmSize": 5, "maxItems": 5000 }

// 2. Licence verification — searches EVERY status, including disbarred and deceased
{ "registrationNumbers": ["1000629", "2222222"] }

// 3. This year's admits — the bar-prep / first-practice-management-tool cohort (4,988 of them)
{ "statuses": ["Currently registered"], "admittedFromYear": 2026, "maxItems": 5000 }

✨ Why use this scraper

• Firm head-count on every row. firmAttorneyCount is computed across the whole currently-registered register before a single row is emitted, so maxFirmSize: 5 hands you solo and small-firm attorneys and leaves BigLaw behind. No other attorney or law-firm scraper on this Store advertises a firm head-count field. Details and the measured normalisation are two sections down.
• It is the register, not a marketing directory. Status, registration number and admission year come from the New York State Unified Court System's own file, published under 22 NYCRR 118.
• Fast, because it is bulk data, not a search form. Other state-bar scrapers drive a name-search page one query at a time and warn that a full-state crawl takes hours. This one asks New York's Socrata API for the slice you want. The prefilled default run returned 100 rows in 9.4-10.9 seconds on Apify, measured over three consecutive runs on 2026-08-14. A 1,000-row Manhattan pull measured 10.7 seconds on the same path, so the cost is the row cap, not the wait.
• The dead people are already filtered out. 49,199 of the 432,690 registrations are deceased attorneys — 11.4% of the file. The statuses prefill removes them before you are billed for one.
• Verification and lead-gen in one Actor. A registration-number lookup deliberately overrides the status prefill so a disbarred or suspended attorney still comes back. A licence check that hides a bad status is worse than no check.
• No wall to beat, and we say so. data.ny.gov is a public open-data API: no key, no login, no anti-bot, no geo gate. Measured 16/16 requests OK direct from an Apify container over 16,000 rows, 0 duplicates. The proxy field defaults to off because a proxy here only adds a failure mode.
• Pagination that cannot silently skip rows. Every request pins $order=registration_number, and the run asserts uniqueness as it walks.

🎯 Use cases

You are a…	Use it to…
Practice-management / legal-tech SaaS (Clio, MyCase, Filevine, Smokeball, Lawmatics class)	Pull maxFirmSize: 5 — solo and small-firm attorneys are the entire ICP, pre-filtered, with a phone number on 100% of the default slice
Malpractice-insurance broker or carrier	Segment by yearsLicensed, county and firm size; renewal timing keys off nextRegistrationDue
CLE and bar-prep marketer	Set admittedFromYear to the current year — 4,988 attorneys were admitted in 2026 so far
Compliance / vendor onboarding	Paste registrationNumbers and get status, admission year and judicial department back for each; disbarred and suspended attorneys are returned, not hidden
Legal recruiter	Filter a county, then cut by lawSchool and yearAdmitted after export
Law-firm M&A and succession advisor	maxFirmSize: 1 plus admittedToYear: 1990 is a list of long-tenured solos with no succession plan
Legal directory / lead marketplace	Seed a New York directory from the authoritative register instead of scraping five commercial ones

📥 Supported inputs

Input	Example	What it does
statuses	["Currently registered"]	Multi-select of the 11 exact New York status strings, each labelled with its live row count
withContactOnly	true	Keep only rows with a phone or street address (251,320 of 293,648)
registrationNumbers	["1000629"]	Licence verification by bar number; searches all statuses
lastNames	["BRODY"]	Exact surname match, case-insensitive (BRODY = 93 registrations)
counties	["New York"]	Multi-select of all 62 NY counties with counts
cities	["BROOKLYN"]	Free text, case-insensitive
states	["NJ"]	Business-address state, multi-select
admittedFromYear / admittedToYear	2026	Admission-year cohort
maxFirmSize	5	Solo / small-firm filter on the computed head-count
maxItems	100	Hard row cap, and your cost ceiling. Raise it for a real pull

Not supported: email addresses (New York publishes none — see Honest limits), states other than New York, disciplinary case documents, fuzzy or prefix name search (lastNames is an exact match), and anything behind a login. This Actor reads one public file and does not touch the courts' own attorney-search web app.

The small-firm filter — the reason to use this over a CSV download

Clio, MyCase, Filevine, Smokeball and Lawmatics do not sell to Kirkland & Ellis, which alone has 1,242 registered New York attorneys. Before it walks a single row this Actor sweeps the whole register grouped by firm and stamps a firmAttorneyCount on every record. Then:

• maxFirmSize: 1 → solo practitioners
• maxFirmSize: 5 → solo and small firms — 115,242 currently-registered attorneys sit at firms of five or fewer (measured 2026-08-12)
• maxFirmSize: 20 → boutiques and below
• empty → everyone, BigLaw included

Attorneys who list no firm at all (typically solos and in-house counsel) are always kept. The head-count is measured across currently-registered NY attorneys statewide, not just within your result slice, so it means the same thing on every row.

Why the head-count is normalised, and why that matters to you

Attorneys type their own firm name, so the raw strings fragment badly:

Raw company_name	Attorneys
KIRKLAND & ELLIS LLP	977
KIRKLAND & ELLIS	174
KIRKLAND AND ELLIS LLP	43
KIRKLAND & ELLIS, LLP	26
KIRKLAND AND ELLIS	12

One firm, five spellings. A naive GROUP BY company_name would tell you a Kirkland partner works somewhere with 12 lawyers and drop them into your small-firm list. Measured on 2026-08-12 across the currently-registered register — the same population the head-count is computed over — grouping raw puts 128,524 attorneys at firms of five or fewer; grouping normalised puts 115,242 there. The naive version mislabels 13,282 BigLaw attorneys (10.3%) as small-firm leads.

So this Actor case-folds, expands &, strips punctuation and strips entity suffixes (LLP/LLC/PLLC/PC/LP/PA/…) before counting — 109,325 raw name strings collapse to 96,549 firms. Placeholders like SELF EMPLOYED, SOLE PRACTITIONER, RETIRED and N/A (816 attorneys) are treated as no firm rather than pooled into one giant pseudo-firm, so those solos stay in your small-firm list where they belong.

The five largest firms in the register after normalisation, if you want to sanity-check it yourself: Kirkland & Ellis 1,242 · Latham & Watkins 1,208 · Simpson Thacher & Bartlett 1,118 · Paul Weiss 983 · Davis Polk & Wardwell 947.

Honest residual: normalisation is string-based, not entity resolution. It will not merge SKADDEN ARPS with SKADDEN, ARPS, SLATE, MEAGHER & FLOM, and a firm whose three attorneys each invented a different abbreviation will still read as three ones. The error that remains is a small under-count at the low end, never the 900-attorney over-inclusion the raw version produces.

📦 Output sample

One real row, exactly as the Actor emitted it on 2026-08-12:

{
  "registrationNumber": "1000322",
  "fullName": "RONALD GABRIEL RUSSO",
  "firstName": "RONALD",
  "middleName": "GABRIEL",
  "lastName": "RUSSO",
  "suffix": null,

  "firmName": "SCHLAM STONE & DOLAN, LLP",
  "firmAttorneyCount": 22,

  "phone": "(917) 207-9328",
  "address": "26 BROADWAY STE 1900",
  "address2": null,
  "city": "NEW YORK",
  "state": "NY",
  "zip": "10004-1703",
  "county": "New York",
  "country": "UNITED STATES OF AMERICA",

  "yearAdmitted": 1974,
  "yearsLicensed": 52,
  "judicialDepartment": 2,
  "lawSchool": "St. John's University School of Law",

  "status": "Currently registered",
  "isCurrentlyRegistered": true,
  "nextRegistrationDue": "2028-03",
  "hasContact": true,

  "sourceUrl": "https://iapps.courts.state.ny.us/attorneyservices/search",
  "scrapedAt": "2026-08-13T01:42:51.424Z"
}

Four fields people misread:

• firmAttorneyCount: null is not zero. It means the attorney listed no firm, or listed a placeholder like RETIRED. Those rows are always kept by maxFirmSize, because an attorney with no firm is the smallest firm there is. A literal 0 is different again: the firm is named, but it has no currently-registered New York attorney left — which is what you see on deceased, resigned and disbarred rows returned by a licence lookup or an all-statuses run.
• firmAttorneyCount counts currently-registered New York attorneys only. A national firm's real head-count is larger; this is its New York bar footprint, which is the number that matters when you are deciding whether it is a small-firm lead.
• nextRegistrationDue is YYYY-MM, not a date. New York publishes a month and a year ("May 2028"); nothing more precise exists.
• Names and firms are returned exactly as New York publishes them — mostly upper case. They are not title-cased, because doing so corrupts McDONALD and O'BRIEN. Case-fold at your end.

🗂 Key output fields

Field	Type	Notes
registrationNumber	string	New York bar number. The dedupe key — unique per attorney
fullName, firstName, middleName, lastName, suffix	string	As published; suffix is filled on 1.8% of rows
firmName	string	Raw string as typed by the attorney
firmAttorneyCount	integer	Computed by this Actor, not published by New York
phone	string	Normalised to (XXX) XXX-XXXX
address, address2, city, state, zip, country	string	Registered business address; ZIP includes +4 when published
county	string	In-state addresses only — 53.9% of the default slice
yearAdmitted, yearsLicensed	integer	yearsLicensed is computed from the current year
judicialDepartment	integer	Appellate Division department of admission, 1–4
lawSchool	string	99.9% filled on the default slice
status, isCurrentlyRegistered	string / boolean	One of the 11 exact New York status strings
nextRegistrationDue	string	YYYY-MM — New York's biennial re-registration month
hasContact	boolean	True when a phone or street address is present
sourceUrl, scrapedAt	string	Provenance. scrapedAt is ISO-8601 UTC

There is no email field, and none can be produced from this source.

Field fill — measured on the 251,320-row default slice

Exact counts over every currently-registered attorney with contact details, measured 2026-08-12 — not a sample, not an estimate:

Field	Fill
name, registration number, admission year, judicial dept., status	100%
phone	100%
nextRegistrationDue	100%
lawSchool	99.9%
address / city	93.1%
firmName	87.9%
firmAttorneyCount	87.9% minus the 816 rows carrying a placeholder
state / zip	84.5%
middleName	79.6%
county	53.9%
address2	8.1%
suffix	1.8%

The number that could mislead you: county is 53.9%, because New York only populates it for in-state addresses — 135,579 of the 251,320 rows in the default slice. On a wider basis it is thinner still: the same 135,579 rows are only 46.2% of all 293,648 currently-registered attorneys, and 163,314 rows carry a county across the whole 432,690-row file. Filtering by county therefore silently restricts you to New York-based attorneys — usually what you want, but choose it deliberately. state/zip trail city by ~9 points for the same reason: a slice of registrants give a foreign or incomplete address, and country is returned so you can drop them.

address2 (8.1%) and suffix (1.8%) are thin because New York genuinely publishes them that rarely, not because they are missed: the registry’s own column statistics report street_2 on 23,439 of 432,758 registrations and suffix on 7,871 — re-measured against the live source on 2026-08-19, and the Actor returns every one of them.

The defaults are worth roughly 3.5× on every contact field. With statuses and withContactOnly cleared, a live run over the raw file returned the first 1,500 rows at 28.2% phone, 24.3% address, 21.5% firm (reproduced 2026-08-12). Every run prints its own measured fill in the log.

Read this before you run it

1. 432,690 registrations is not 432,690 leads. 49,199 of them — 11.4% of the file — are deceased attorneys, and another 49,428 are delinquent, suspended, resigned, disbarred or incapacitated.

   Status	Rows
   Currently registered	293,648
   Deceased	49,199
   Due to reregister within 30 days of birthday	40,415
   Suspended, delinquent	19,097
   Delinquent	16,492
   Resigned	7,136
   Disbarred	2,190
   Incapacitated	1,904
   Suspended, currently registered	1,405
   Resigned from bar - disciplinary reason	718
   Suspended, due to reregister	486

2. A registration-number lookup ignores your status and contact filters, on purpose. Verifying a licence under the Currently registered prefill would hide exactly the attorneys you are checking for. The run logs that it switched modes.

3. county restricts you to New York-based attorneys. See the fill table above.

4. One row per registration, and a registration number belongs to one attorney. The Actor deduplicates on it while it walks and warns if it ever sees a repeat.

5. A registration number that does not exist fails the run rather than returning an empty dataset — with a message naming the number. That is the verification answer, and it costs you nothing because no row was delivered.

💰 Pricing

$0.003 per attorney returned — $3.00 per 1,000. Charged on the attorney-scraped event. No monthly platform fee from this Actor.

Run	Attorneys	Cost
Default prefilled run (Manhattan, currently registered)	100	$0.30
1,000 Manhattan attorneys (maxItems: 1000)	1,000	$3.00
Everyone admitted in 2026	4,988	$14.96
Every currently-registered attorney in Manhattan	77,761	$233.28
Every currently-registered attorney with contact details, statewide	251,320	$753.96
The entire register, every status including the deceased	432,690	$1,298.07

Rows are charged as they are pushed — hit a budget cap and you get whole rows and stop, never a half-billed dataset. Verified on the run behind this page: 1,000 rows delivered, exactly 1,000 attorney-scraped events charged. You are never charged for a row your filters removed: statuses, counties, cities, states and the admission-year cohort are applied server-side by New York, so those rows are never even fetched, and maxFirmSize drops rows before they reach the dataset. You are never charged twice for the same registration number. maxItems is your hard cost cap (default 100; there is no unbounded mode — set it to the number you want).

Honest limits

• No email addresses. New York publishes none, for any attorney. Name, firm, postal address, phone, county, law school, admission year and status are what 22 NYCRR 118 requires and all that exists in this file. There is no email field and nothing here can produce one. Phone (100% of the default slice) and postal mail (93.1%) are the channels. Anyone selling you "New York bar emails" generated them somewhere else.
• New York only. Not a national dataset, and this page does not imply it is. See the FAQ for why Florida was evaluated and rejected.
• county is 53.9% of the default slice, and only exists for in-state addresses.
• state/zip are 84.5%, firmName 87.9%. A missing firm means the attorney did not file one, not a scraping failure — the field is self-reported on the biennial registration form.
• Firm normalisation is string matching, not entity resolution. It merges

  KIRKLAND AND ELLIS LLP
  into KIRKLAND AND ELLIS; it does not merge SKADDEN ARPS with the full firm name. The residual error is a small under-count at the low end.
• Freshness is the publisher's, not ours. The NYS Office of Court Administration's own metadata states a posting frequency of Quarterly; the file's rowsUpdatedAt read 2026-08-12T05:03:58Z when this page was written, i.e. it had in fact been refreshed that morning. Every run logs the live timestamp so you can see the true age of the data you bought. Registration data is also inherently biennial — attorneys update their address when they re-register, not when they move.
• lastNames is an exact match, not fuzzy. SMITH will not return SMITHERS and a misspelled surname returns nothing. The run fails loudly rather than shipping an empty dataset.
• No login, no API key, no CAPTCHA-solving service, and no attempt to touch the courts' own attorney-search web app. This Actor reads one public open-data file.

⚙️ How it works (technical)

1. Reads https://data.ny.gov/resource/eqw2-r5nb.json — New York's public Socrata endpoint for the NYS Attorney Registrations dataset. No key, no auth, no browser.
2. Translates your inputs into one SoQL $where clause, so status, county, city, state and admission-year filtering all happen on New York's servers. A county × status slice costs a handful of requests instead of a 432,690-row download.
3. Asks for the match count first and refuses to start if it is zero, with a message naming what you asked for.
4. Sweeps company_name grouped by count (22 paged requests), normalises each name and builds the firm head-count map, then asserts the map looks sane — fewer than 1,000 distinct firms means the schema changed upstream and the run stops rather than shipping wrong head-counts.
5. Walks the result 5,000 rows at a time with $order=registration_number pinned. Socrata does not guarantee stable ordering across pages without it; without the pin, offset paging silently repeats and skips rows.
6. Normalises each record (phone to (XXX) XXX-XXXX, "May 2028" to 2028-05), checks the registration number against the seen-set, applies maxFirmSize, and pushes in batches of 1,000 with the charge attached to the push.

Transport, measured through Apify: direct, no proxy, 16/16 requests OK (100%) over 16,000 rows, 0 duplicates. There is no anti-bot, no geo gate and no rate limit worth dodging on data.ny.gov, which is why the proxy input defaults to off — the sibling Texas licence actor died on a proxy 590 while the direct request succeeded. Four runs on 2026-08-12/13 (1,000 rows on Apify, plus 1,000, 1,500 and a 93-row surname query locally) logged zero duplicate registration numbers between them.

When a run fails

• Zero matching rows → the run throws with the filters you asked for and a reminder of the exact New York strings. It never green-lights an empty dataset.
• A registration number not in the register → throws naming the number. That is the answer.
• The firm sweep returns fewer than 1,000 firms → throws, on the assumption that New York changed the schema, rather than emitting rows with wrong head-counts.
• Duplicate registration numbers → skipped, counted and warned about in the log.
• Requests retry four times with backoff before the run gives up.

❓ FAQ

Can I get attorney email addresses? No. New York publishes none, for any attorney, in this register — the file has 20 columns and email is not one of them. You get phone on 100% of the default slice and a postal address on 93.1%.

Is this the same as a bar-number lookup? Yes, and it is a first-class input. Paste registration numbers into registrationNumbers and you get one row each with status, admission year, judicial department, firm and address. It searches every status, so a disbarred or deceased attorney is returned rather than silently dropped.

Can I verify a specific attorney by name? Set lastNames — an exact, case-insensitive surname match (BRODY returns 93 registrations across all statuses). Clear statuses and withContactOnly first, or you will only see the currently-registered ones with contact details. Filter to the right first name after export.

How current is the data? The NYS Office of Court Administration states a quarterly posting frequency; the file's own rowsUpdatedAt timestamp read 2026-08-12T05:03:58Z when this page was written. Every run logs that timestamp, so you always know how old the data you just bought is.

Why New York only? Because New York is the second-largest bar in the United States — 293,648 currently-registered attorneys — and it is the one that publishes the whole register as a clean bulk open-data feed, which is why this Actor can hand you 1,000 rows in seconds instead of hours of form-driving. Want another state? Ask on the Issues tab. Florida was evaluated as the second source and rejected: The Florida Bar's member profiles sit behind a Cloudflare interstitial (measured: HTTP 429 "Just a moment…" on every profile fetch, including with a warmed session cookie), and its records carry a different field set — bolting that on would mean shipping half-empty rows under a New York schema.

Do I need a proxy? No. data.ny.gov is a public open-data API with no anti-bot and no geo gate; measured 16/16 requests OK direct. The proxy field is there only if your account needs a fixed egress IP, and turning it on adds a failure mode rather than removing one.

Will I get duplicates? What about across two runs? Not within a run: the Actor pins $order=registration_number and checks each number against a seen-set before pushing, so you are never billed twice for the same attorney. Across runs, dedupe on registrationNumber at your end — it is stable and unique per attorney.

How many rows can I get in one run? As many as maxItems allows; the whole default slice is 251,320 attorneys and the whole file is 432,690. There is no server-side result cap to work around — this is bulk open data, not a search form that stops at 500.

Do I get charged for rows that get filtered out? No. Status, geography and admission-year filters run on New York's servers so those rows are never fetched, and maxFirmSize drops rows before the dataset write. You are billed per attorney delivered, on push.

Will a run ever succeed with zero rows? No. Zero matches throws with the scope you asked for. A green run with an empty dataset would be a bug, not a result.

Can I schedule it? Yes — use Apify Schedules. Registration data changes slowly (attorneys re-register biennially), so a monthly or quarterly refresh is usually enough; check the freshness timestamp the run logs.

Something looks wrong — how do I debug it? Read the run log. It prints the match count before it starts, the firm-sweep totals, the emitted count per page, the upstream freshness timestamp, and the measured field fill for exactly the rows you received. If those disagree with this page, open an issue.

💬 Support & 🔎 more scrapers

Missing a field, want another state, or need a filter that is not here? Open an issue on the Issues tab of this Actor.

Actor	What it is	Why you would use it instead
California State Bar Attorney Scraper	The California register — phone, firm, licence status	You need California, the largest US bar, from its own official source
This Actor	The New York register, plus a computed firm head-count	You need New York, bulk, with solo/small-firm targeting or a bar-number check

We also build directory-side legal scrapers (Martindale-Hubbell peer ratings, Justia, Super Lawyers) — see our Apify profile for what is currently live. Those cover marketing directories with ratings and practice areas; this Actor covers the licence register. If you need practice areas or peer ratings, a directory scraper is the right tool; if you need licence status and a bar number you can trust, this one is.

⚠️ Disclaimer

This Actor is an independent tool and is not affiliated with, endorsed by, or sponsored by the New York State Unified Court System, the New York State Office of Court Administration, the New York State Bar Association, or any of their affiliates. All trademarks are the property of their respective owners.

It reads a single public open-data file published by the NYS Unified Court System at data.ny.gov under 22 NYCRR 118. It does not log in, does not solve CAPTCHAs, does not touch authenticated or member-only endpoints, and collects no data that New York does not publish for public licence-verification use.

Rows describe identifiable people. You are responsible for complying with New York's terms and with how you use the data — including GDPR/CCPA where applicable, CAN-SPAM and telemarketing rules for outreach, and the New York Rules of Professional Conduct on attorney advertising and solicitation (Rules 7.1–7.3) if you are a law firm marketing to or through this list.

SEO Keywords

new york state bar scraper, ny attorney registration search, new york attorney lookup, attorney license verification, bar number lookup, lawyer license check api, attorney status check, ny bar number search, new york lawyer directory scraper, attorney license lookup new york, nys attorney registration data, lawyer lead generation new york, solo practitioner list, small law firm leads, attorney contact list ny, law firm size data, legal lead generation scraper, attorney verification api, disbarred attorney check, ny courts attorney search, new york bar admission year data, law school alumni attorney list, attorney phone number list, legal tech icp data, malpractice insurance leads, cle marketing list, attorney public records new york