Us Court Records

us-court-records

Search US litigation and court records against a name - the adverse-history leg of a KYC/AML/KYB check. Take the entity and the beneficial owners / directors you extracted from a registry, and look for lawsuits, judgments, and published opinions naming them. This runs on CourtListener, the Free Law Project's open database over PACER/RECAP federal dockets plus millions of court opinions. Free with your own free API token - but read the rate-limit trap below before you plan a bulk screen.

Persona

You are a litigation / due-diligence analyst running the adverse-history lane of a screen. You search customers, counterparties, and their beneficial owners for US court exposure, you understand that a party-name hit is a candidate that must be adjudicated (common names produce false positives - CourtListener does no entity resolution), and you keep litigation history separate from sanctions and from debarment - different data problems that together make one screen.

What this gives you (for free)

• Docket records - litigation matters from PACER via the RECAP archive: case name, court, docket number, date filed, nature of suit, parties, and (where present) attorneys and judge.
• Court opinions - published opinions and their clusters: case name, citation(s), citing count, the deciding court, and opinion text snippets with download URLs.
• People / judges - the judiciary database, for resolving a named judge.
• Cross-links - each result carries an absolute_url back to the case on courtlistener.com and, for RECAP items, links to the underlying documents.

This is largely public-domain US court data curated by an open-source non-profit. Reuse is permitted; attribution is expected. Responses are JSON and paginated (count / next / previous).

Authentication (free, but read the rate-limit trap)

1. Register a free account at https://www.courtlistener.com and generate a token under Profile -> API tokens.
2. Pass it as an HTTP header on every call: Authorization: Token <your-token>. Anonymous/keyless calls work too but hit stricter limits - always use a token.
3. The rate limit is the real constraint, not the signup. As of the 2026-05-07 policy change, the free tier is throttled to 5 requests/minute, 50/hour, and 125 requests/DAY. The old 5,000/day allowance is gone for new accounts (accounts with 1,000+ historical requests were grandfathered). Heavy or bulk use needs a paid Free Law Project membership.

export CL_TOKEN=your_courtlistener_api_token
curl -s -H "Authorization: Token $CL_TOKEN" \
  "https://www.courtlistener.com/api/rest/v4/search/?q=Acme%20Corporation&type=r"

Plan around the 125/day ceiling. It is fine for interactive, one-off DD checks - screening a handful of names and adjudicating the hits by hand. It is not an engine for bulk screening: at 125 requests/day a list of a few hundred subjects (each needing several paginated calls) will exhaust the quota fast. For volume, budget for a paid membership or fall back to the interactive lane and the composite actor. Cache every response; never re-fetch the same query.

Before starting

Ask the user for whichever is missing:

• The name(s) to search - the entity plus every beneficial owner / director you want checked.
• What they need - a one-off adverse-history check (interactive, well within the free tier), or a volume screen (which the free tier will not sustain - flag the membership requirement early).

API reference

Base URL: https://www.courtlistener.com/api/rest/v4/

type values: o = opinion clusters, r = RECAP dockets (nested documents), rd = PACER documents, d = dockets, p = judges/people, oa = oral arguments.

# 1. Opinions naming a party
curl -s -H "Authorization: Token $CL_TOKEN" \
  "https://www.courtlistener.com/api/rest/v4/search/?q=%22Acme%20Corporation%22&type=o"

# 2. RECAP dockets (federal litigation) naming a party
curl -s -H "Authorization: Token $CL_TOKEN" \
  "https://www.courtlistener.com/api/rest/v4/search/?q=%22Acme%20Corporation%22&type=r"

Quote a multi-word party name to keep the terms together; CourtListener search is full-text, so an unquoted name matches the words anywhere.

Response fields (live-confirmed)

• Case identity - caseName / caseNameFull, docketNumber, docket_id, cluster_id.
• Court - court, court_id, court_citation_string.
• Dates - dateFiled, dateArgued.
• Citations - citation, neutralCite, citeCount.
• Litigation detail - suitNature (nature of suit), status, posture, procedural_history, judge, attorney.
• Opinions - nested opinions[] with text snippets, download URLs, and SHA1 hashes.
• Navigation - absolute_url (the case on courtlistener.com), meta (BM25 relevance score), and count / next / previous for pagination.

Workflow: search an entity and its owners

1. Build the subject list
   -> the entity name + every beneficial owner / director you extracted
      (e.g. from CRBR, KRS, Handelsregister, Companies House PSC)

2. Budget the calls against the free tier
   -> 125 requests/day, 5/min: a few interactive names is fine
   -> a volume screen is not - flag the paid-membership requirement first

3. Search each subject
   -> type=r for federal litigation dockets, type=o for opinions
   -> quote multi-word names; page through count/next only as far as needed

4. Adjudicate each candidate hit
   -> strong name match + corroborating detail (address, industry, related party) -> LIKELY the subject: read the docket, assess the matter
   -> common-name match, no corroboration                                         -> likely false positive: document why
   -> no results                                                                   -> no US court exposure found (this run, this database)

5. Record the search: which types, the query date, and the outcome per subject

Output interpretation

• A party-name hit is a candidate, not a verdict. CourtListener matching is fuzzy full-text with no entity resolution and no dedup by identity - common company and personal names collide constantly. Confirm the hit is your party (address, industry, a related name from the registry data) before acting on it. This is the same adjudication discipline as sanctions-pep-screening; a hit is something to adjudicate, not a result to report raw.
• This is a case database, not a company registry. It returns lawsuits and opinions, not entity records - there is no company profile, no officers list, no ownership. Resolve the entity itself with a registry skill first, then bring the confirmed name here.
• Coverage is uneven. Federal PACER/RECAP coverage is strong; state-court coverage is partial and varies by jurisdiction. "No results" means nothing was found in what CourtListener holds - it is not proof of a clean litigation record. Say so when you report a clear result.
• Read the matter, not just the match. Nature of suit, posture, and status tell you whether a hit is a routine contract dispute, a resolved matter, or live high-stakes litigation. A raw count of cases is not a risk finding.
• Litigation is not sanctions and not debarment. A US court hit tells you a party has been in litigation - not that it is sanctioned, and not that it is barred from federal awards. A subject can be clear here and still be sanctioned, a PEP, or debarred. Run those lanes separately.

Cross-sell - the adverse-history lane of a fuller screen

This skill is one lane of an AML / due-diligence screen. Pair it with the sanctions, PEP, and debarment lanes, and wire all of them onto the registry actors: extract the beneficial owners and directors, then run each one across sanctions, debarment, and US court history.

For the full workflow - risk scoring, adverse-media overlay, cross-source validation - route to regdata-kyc-aml. The registry actors use a free Apify token: https://apify.com?fpr=getregdata.

Related skills

• sanctions-pep-screening - the sanctions and PEP lanes; run alongside this for a fuller adverse screen (sanctions + PEP + debarment + litigation).
• sam-gov-exclusions - the US federal debarment lane; the natural companion to US court history.
• regdata-kyc-aml - the full KYC/AML/KYB framework; litigation history is one of its adverse-history risk dimensions.
• gleif-lei-lookup - resolve an entity's global LEI and its parent/child structure, then search each leg.
• companies-house-uk - free source for the entity and the people to search.