FINRA Arbitration Awards Scraper

FINRA Arbitration Awards Scraper - Awards Online Search

Search and download FINRA's Arbitration Awards Online archive: 66,600 securities arbitration awards, orders and motions from 1988 to this week. Every row carries the case number, date of award, forum, claimants, respondents, both sides' representatives, the arbitrators, the hearing site and the link to the award PDF. No login, no API key, no browser. Export to CSV, JSON, Excel, or XML.

FINRA's own Awards Online search shows fifteen results at a time, offers no export, and stops at four filters. This Actor reads the same public index, applies your filters, and then goes one step further: it opens the award PDF and parses what the index never exposes, including the damages awarded, the forum fees, the arbitrator roles, the causes of action and the expungement findings with their CRD numbers.

Who uses it	What they use FINRA award data for
Securities and investor-rights lawyers	Award history for a firm, a broker or an arbitrator before filing or picking a panel
Compliance and legal teams at broker-dealers	Every award naming the firm, with the amounts, fees and causes of action
Expert witnesses and damages consultants	Comparable awards by cause of action, hearing site and year for a damages opinion
Financial journalists and academics	A structured corpus of 66,600 arbitration outcomes going back to 1988
Legal tech and data vendors	A normalised feed of new awards with stable case ids and PDF links

What it does

FINRA publishes every arbitration award it administers, plus the court orders and motions filed against them. The archive holds 66,617 documents: 65,372 awards, 475 orders to confirm, 449 orders to vacate, 130 motions to vacate, 27 motions to confirm and 175 other documents, across eight forums (NASD 36,421, FINRA 25,449, NYSE 4,130, CBOE 197, AMEX 147, MSRB 139, Pacific Exchange/ARCA 78, PHLX 56).

Every row returns:

• 🧾 The index record: case number, document title, date of award, award year, the year the case was filed, document type, forum, panel composition, hearing site split into city and state, and the direct PDF link.
• 👥 The parties: claimants and respondents as clean lists with counts, each side's representatives, and whether each side was represented at all.
• ⚖️ The panel: the arbitrators who signed, as a list and as text, with the count.
• 🚩 Flags: whether FINRA has expunged the case from CRD registration records, whether the case has other documents on file, and whether the index carries a summary.
• 💵 Damages (optional): compensatory, punitive, attorney fees, expert fees and costs awarded, the total, the interest rate and its start date, the amount the claimant asked for, and whether the claims were denied.
• 🧾 Forum fees (optional): filing fees, member surcharge and process fee, total hearing session fees, the session count, and each assessment with the party it was charged to.
• 🧑‍⚖️ Arbitrator roles (optional): each arbitrator as public, non-public, presiding chairperson or sole arbitrator, with the panel size and the chairperson.
• 🔎 Expungement (optional): requested, granted or denied, the CRD occurrence numbers, the registered persons with their CRD numbers, the Rule 2080 affirmative findings, and whether court confirmation is required.
• 📄 The document (optional): counsel and law firms, the causes of action, the case timeline, the narrative sections, and the full award text.

searchTerms is verified against the award itself. With a date range, every award in the range is downloaded and checked for the term, so churning or Rule 12206 finds the awards that really say it. Without a date range FINRA's index supplies candidates, which are checked the same way and dropped when the term is not in the document. FINRA's index alone only searches names, case numbers and, for awards before July 2007, the award text.

Every row contains: caseId, documentTitle, dateOfAward, awardYear, caseFiledYear, documentType, forum, panelComposition, hearingSite, hearingCity, hearingState, claimants, claimantsText, claimantCount, respondents, respondentsText, respondentCount, claimantRepresentatives, respondentRepresentatives, claimantRepresented, respondentRepresented, neutrals, neutralsText, arbitratorCount, summary, hasSummary, hasRelatedCases, isExpungedInCrd, documentUrl, caseUrl, indexId and scrapedAt.

What you can do with FINRA arbitration award data

Build a firm's award history.

Set partyNameContains to a broker-dealer and a date range, turn on damages, and you get every award naming that firm with the compensatory, punitive and attorney-fee amounts, the causes of action and the hearing site. That is the table a plaintiff's lawyer, a compliance officer or a journalist needs and that no public search returns.

Research an arbitrator before you strike the list.

arbitratorNameContains keeps only awards decided by that arbitrator. With the panel block on, each row says whether they chaired, whether they sat as a public or non-public arbitrator, what the panel awarded and how the fees were assessed.

Track expungements.

Turn on the expungement block and filter with onlyExpunged or panelCompositions: ["Special Expungement Panel"]. Each row gives the occurrence numbers, the registered persons with their CRD numbers, the Rule 2080 finding the panel made and whether court confirmation is still required.

Benchmark damages by cause of action.

Search a phrase such as unsuitability or selling away over a date range, keep the damages and causes-of-action blocks, and you have the awarded-versus-requested ratio by year, forum and hearing site.

Why choose this scraper

	What you get
The whole archive	66,600 documents from 1988 to this week, including the NASD and exchange forums that predate FINRA
What the index hides	Damages, fees, arbitrator roles, counsel, causes of action, case timeline and expungement findings parsed from the award PDF itself
Filters that are measured, not guessed	Forum, document type and panel composition run inside FINRA's index; every value was checked against the archive's own totals
Verified text search	Every search hit is checked against the award PDF; an award that does not contain your term is never returned or charged
Honest empties	Awards published as scanned images say so instead of returning blank damages, and are not charged
Clean parties	Claimants, respondents, representatives and arbitrators split into lists that keep Drexel Hamilton, LLC in one piece
No login, no browser, no key	Plain HTTP against public endpoints; 900 index rows in 23 seconds

How it compares

Nobody else on the Apify Store sells this archive. The closest listings are BrokerCheck scrapers, which return a broker's disclosure summary rather than the award, and generic court-record actors that do not cover FINRA arbitration at all. Against FINRA's own site the difference is exports, filters and the parsed PDF.

Feature	ParseForge	FINRA Awards Online	BrokerCheck actors	Court-record actors
Bulk export (CSV, JSON, Excel)	Yes	No	Yes	Yes
Awards back to 1988, all eight forums	Yes	Yes	No	No
Damages and fees parsed from the award	Yes	No	No	No
Arbitrator roles and panel composition	Yes	Partial	No	No
Expungement findings with CRD numbers	Yes	No	No	No
Text search verified against the award	Yes	No (fuzzy index hits)	No	Sometimes
Price per row	$0.007	Free, manual	$0.0015 to $0.05	$0.003 to $0.005

What an award record looks like

{
  "caseId": "26-00668",
  "documentTitle": "26-00668.pdf",
  "dateOfAward": "2026-08-26",
  "awardYear": 2026,
  "caseFiledYear": 2026,
  "documentType": "Award",
  "forum": "FINRA",
  "panelComposition": "N/A",
  "hearingSite": "Houston, TX",
  "hearingCity": "Houston",
  "hearingState": "TX",
  "claimants": ["Merrill Lynch Pierce Fenner & Smith Inc."],
  "claimantsText": "Merrill Lynch Pierce Fenner & Smith Inc.",
  "claimantCount": 1,
  "respondents": ["Ignacio Hawley"],
  "respondentsText": "Ignacio Hawley",
  "respondentCount": 1,
  "claimantRepresentatives": ["Maria V. Martin"],
  "respondentRepresentatives": ["Ignacio Javier Hawley"],
  "claimantRepresented": "Yes",
  "respondentRepresented": "Yes",
  "neutrals": ["Maryanne M. Esser"],
  "neutralsText": "Maryanne M. Esser",
  "arbitratorCount": 1,
  "summary": "Not Disclosed",
  "hasSummary": "No",
  "hasRelatedCases": "No",
  "isExpungedInCrd": "No",
  "documentUrl": "https://data-portal.finra.org/aao_documents/26-00668.pdf",
  "caseUrl": "https://www.finra.org/arbitration-mediation/arbitration-awards-online?aao_radios=case_id&field_case_id_text=26-00668",
  "indexId": "26-00668_5e6cd9fd",
  "scrapedAt": "2026-08-27T16:21:11.532Z"
}

With the document blocks on, the same row also carries compensatoryDamagesUsd, punitiveDamagesUsd, attorneyFeesUsd, awardedCostsUsd, totalAwardedUsd, requestedCompensatoryUsd, interestRate, feeItems, feeAssessments, arbitratorPanel, chairperson, claimantCounsel, lawFirms, causesOfAction, natureOfDispute, hearingDates, expungementPersons and the narrative sections. A real award from March 2024 came back with compensatoryDamagesUsd: 718178, punitiveDamagesUsd: 200000, attorneyFeesUsd: 417453, interestRate: "9%" and causesOfAction listing seven claims.

Configure the run

Leave everything empty and the Actor returns the most recent awards. Add search terms, case numbers or filters to narrow it, and tick the document blocks you want parsed. summary, panelComposition and hearingSite come back as Not Disclosed or N/A when FINRA does not publish them, which for pre-2000 awards is most of the time.

The most recent 100 awards:

{ "maxItems": 100 }

Every 2024 award naming a firm, with the damages and the panel:

{
  "partyNameContains": "Morgan Stanley",
  "dateFrom": "2024-01-01",
  "dateTo": "2024-12-31",
  "documentTypes": ["Award"],
  "includeAwardDocument": true,
  "includeMonetaryTerms": true,
  "includeArbitratorPanel": true,
  "maxItems": 500
}

Expungement decisions with the CRD numbers and the Rule 2080 findings:

{
  "panelCompositions": ["Special Expungement Panel"],
  "dateFrom": "2023-01-01",
  "includeAwardDocument": true,
  "includeExpungementDetail": true,
  "includeCaseTimeline": true,
  "maxItems": 250
}

Every 2024 award whose text contains a term (each award in the range is downloaded and checked):

{ "searchTerms": ["churning"], "dateFrom": "2024-01-01", "dateTo": "2024-12-31", "documentTypes": ["Award"], "maxItems": 300 }

Free users

Free-plan runs return up to 10 awards as a preview. Upgrade your Apify plan to pull the whole archive.

Run it

1. Create a free Apify account with $5 in credit.
2. Open the FINRA Arbitration Awards Scraper.
3. Add search terms, case numbers or filters, tick the document blocks you want, then click Start.
4. Export the results as CSV, Excel, JSON, or XML from the Dataset tab, or schedule the run weekly to catch new awards.

Run it programmatically through the Apify API (run-sync-get-dataset-items) or the ApifyClient for JavaScript and Python.

Use with AI agents (MCP)

Give an AI agent live access to FINRA's award archive through the Model Context Protocol. Add the Actor to Claude, Cursor, or any MCP client:

claude mcp add --transport http apify "https://mcp.apify.com?tools=parseforge/finra-arbitration-awards-scraper"

Then ask: "What did FINRA panels award against Interactive Brokers in 2024, and how much was requested?", "List the expungement awards from 2025 with their CRD occurrence numbers", or "Which arbitrators chaired the largest awards last year?"

Troubleshooting

Why am I getting no results?

A filter combination can be empty: panelCompositions tags only 6,000 of the 66,600 documents, and documentTypes: ["Motion to Confirm"] matches 27 in the whole archive. Drop one filter at a time. Case numbers must be exact, in FINRA's own format (24-00627, 2007-016751).

Why fewer rows than I asked for?

Your date range, filters and local filters (partyNameContains, arbitratorNameContains, hearingSites) may simply match fewer awards than maxItems. The run log prints how many each search chain returned.

Why does a search term return fewer awards than FINRA's website?

FINRA's own search is fuzzy: 12206 returns awards that never mention Rule 12206, because for awards issued by FINRA (July 2007 onwards) its index only holds the parties, arbitrators and case number. This Actor checks each award's text and drops those, so the count is smaller and every row is a real hit. Give a date range (dateFrom and dateTo) and every award in it is checked, which is the only way to find a term inside a recent award; without one the Actor can only verify the candidates FINRA's index hands out, and the log says so.

Why is a search run slower than a browse run?

Every award in the range is downloaded (about 100 KB) and parsed to check for your term, roughly 0.7 seconds per award at the default 512 MB (577 awards took 6.5 minutes). About 1,400 awards fit in the default 20-minute timeout; raise the run timeout for a larger range, up to 4,000 awards per run.

Why is a damages field zero?

Because the panel awarded nothing under that head. Expungement-only awards decide no money at all and come back with claimsDenied: "N/A"; awards that denied the claims come back claimsDenied: "Yes". Both are the award's real content.

Why does an award have no parsed sections?

About 7% of the archive is a scanned image with no text layer, including a few recent ones. Those rows say documentTextStatus: "Scanned image with no machine-readable text", keep the index fields, and are not charged for the document or any extraction.

Why is the run slow?

The search index hands out one page of 15 results at a time and rate limits bursts, so a wide date range is walked month by month across several connections. 900 index rows take about 23 seconds; parsing PDFs adds roughly one second per award. Raise the run's memory to give the PDF parser more CPU.

FAQ

Question	Answer
Is this data public?	Yes. FINRA publishes every award in Arbitration Awards Online for anyone to read, and this Actor reads only that public index and those public PDFs.
How far back does it go?	To 1988, when the NASD forum begins. NASD covers 36,421 documents, FINRA 25,449, and the exchange forums (NYSE, CBOE, AMEX, MSRB, Pacific Exchange/ARCA, PHLX) the rest.
How current is it?	Awards appear in the index within days of service. The archive returned an award dated the previous business day during testing.
Can I search the text of the awards?	Yes, and every hit is verified. With a date range the Actor downloads every award in it and keeps only those whose text contains the term. Without a date range it relies on FINRA's index, which for awards after July 2007 only covers parties, arbitrators and case numbers; those hits are still checked against the PDF before they are returned or charged.
Does it give me the award amounts?	Yes, with includeMonetaryTerms. Compensatory, punitive, attorney fees, expert fees, costs, the total, the interest rate and the amount originally requested.
Can I filter by arbitrator?	Yes. arbitratorNameContains keeps only awards that arbitrator decided, and includeArbitratorPanel adds their role on the panel.
What is an expunged award?	One FINRA has removed from a broker's CRD record. The index flags 1,350 of them, and includeExpungementDetail returns the occurrence numbers, CRD numbers and the Rule 2080 findings behind the ruling.
Why do some cases return two rows?	FINRA sometimes files more than one document under a case, such as a corrected award or a motion to vacate. Each document is its own row; hasRelatedCases marks them and includeRelatedCases lists the siblings.
How many awards per run?	Free plan: 10. Paid: up to 1,000,000, bounded by what your filters match.
Is this an official FINRA product?	No. It is unofficial and reads only public FINRA data.

🆘 Need help? Email parseforge@protonmail.com with your run ID, your input, and what you expected.

⚠️ Disclaimer. This Actor is unofficial and is not affiliated with, endorsed by, or sponsored by FINRA. It collects only publicly available data from FINRA's Arbitration Awards Online. Awards name real people, so you are responsible for using the data in compliance with applicable laws, including GDPR, CCPA, and PIPL, and with FINRA's terms. Do not use it to identify, profile, or target individuals.