Enforcement Record Profiler — Company Violations & Lawsuits

Checking whether a company has a record of regulatory or legal trouble means searching a dozen government sites by hand — each with its own quirks — and stitching the results together. Enforcement Record Profiler does it in one call: name an organization, get back a single cited report of its enforcement and accountability record across federal agencies and the courts, with every finding linked to its official source.

What this does

Submit one organization name. The Actor resolves it to a canonical identity, searches the official public-record sources below, and returns a structured, citation-grounded profile.

• Entity resolution — maps the name you give to a canonical identity (official name, known aliases, and SEC CIK + stock ticker when the entity is a public registrant), so findings attach to the right company.
• Multi-source search across official records:
  • Federal courts — dockets and opinions naming the entity (via CourtListener/RECAP).
  • U.S. Department of Justice (DOJ) — press-released charges, indictments, pleas, settlements, and judgments.
  • U.S. Securities and Exchange Commission (SEC) — recent enforcement actions: litigation releases, administrative proceedings, and accounting-and-auditing enforcement. Covers the agency's current enforcement window rather than its full case history.
  • Federal Trade Commission (FTC) — consumer-protection, deceptive/unfair-practice, antitrust, and privacy cases.
  • OSHA — workplace-safety inspections and citations.
  • National Labor Relations Board (NLRB) — unfair-labor-practice charges and representation/election cases naming the entity.
  • Environmental Protection Agency (EPA) — federal environmental enforcement cases (Clean Air Act, Clean Water Act, RCRA, CERCLA/Superfund, TSCA, and more) naming the entity, with penalties and cleanup cost-recovery amounts.
• Per-claim citations — every finding carries the source agency, a title, the date, a penalty/settlement amount when shown, and the exact source URL. Nothing is asserted without a link.
• Neutral summary — a plain-language synthesis of the record across sources, describing what the documents state (not conclusions about guilt).
• Source-coverage audit — for each source: whether it was queried, how many findings it produced, and a note if it was unavailable.
• Three outputs per run — a machine-readable dataset record, a human-readable Markdown report (OUTPUT in the run's Key-Value Store), and the raw findings as JSON (RAW_FINDINGS).

Common workflows this enables:

• A journalist building a sourced timeline of a company's misconduct before filing a story.
• A vendor-risk or ESG analyst screening a supplier or portfolio company for enforcement history.
• A litigation-support researcher pulling the federal cases and agency actions tied to an entity.
• An academic or watchdog compiling accountability data across many organizations.
• An activist investor assembling the regulatory record ahead of an engagement.

Why cited, multi-source profiling matters

The signal an analyst or reporter actually needs — "what has this organization been held accountable for, and where is the proof?" — is scattered across agencies that each publish differently. Doing it by hand is slow and easy to get wrong: miss one agency and you miss the story.

The worse failure mode is the tempting shortcut: ask a general-purpose model "what has Company X done wrong?" and get a fluent paragraph that invents a settlement, misstates a penalty, or cites a URL that doesn't exist. For accountability work, an unverifiable claim is worse than no claim — it's a correction or a libel risk.

This Actor is built the opposite way. It fetches the official agency and court pages, extracts only what is present on them, and attaches the real source link to every finding. The summary is generated from those cited findings, not from model memory. If a source has nothing, the report says so rather than filling the gap. You get something you can publish or hand to counsel, because every line traces back to a government record.

How it compares

Approach	Multi-agency in one pass	Per-claim citations	Grounded in official records	Agent-callable
Manual search across agency sites	✗ (one at a time)	✓ (you copy them)	✓	✗
General web-search LLM	partial	✗ (often fabricated)	✗	partial
Subscription compliance/diligence suites	✓	✓	✓	✗ (seat-licensed)
Enforcement Record Profiler	✓	✓	✓	✓

The distinguishing combination is unified, cited, official-source-grounded, and callable by an agent or a script per profile — without a seat license or a manual multi-site slog.

Input

Field	Type	Required	Default	Description
entity	string	✓	—	Legal name of the organization to profile (one entity per run). Add a parenthetical hint if the name is ambiguous, e.g. "Acme Corp (Delaware)" or a ticker. Organizations and public figures acting in an official capacity only — not private individuals.
sources	array of strings		all	Subset of sources to query. Valid values: courts, doj, sec, ftc, osha, nlrb, epa. Omit to query all. Invalid values are rejected with a clear error.
sinceYear	integer		—	Only include findings dated on or after this year. Omit for no lower bound. Narrowing reduces noise and cost.
maxRecordsPerSource	integer		25	Cap on findings returned per source (1–200). Bounds run cost and filters low-signal volume.

Output

One dataset record per run, plus a Markdown report at OUTPUT and raw findings at RAW_FINDINGS in the run's Key-Value Store.

{
  "entity": "Wells Fargo & Company",
  "resolved_identity": {
    "name": "WELLS FARGO & COMPANY/MN",
    "aliases": ["Wells Fargo & Company"],
    "cik": "0000072971",
    "ticker": "WFC",
    "subsidiaries": []
  },
  "summary": "Across the queried records, federal regulators and courts show multiple actions, including consumer-protection settlements and labor and safety matters. See findings for sourced detail.",
  "total_findings": 23,
  "findings": [
    {
      "source": "doj",
      "title": "Company agrees to civil settlement over billing practices",
      "date": "2020-02-21",
      "category": "criminal/civil",
      "summary": "Settled civil matter; the company agreed to a payment without admitting liability.",
      "amount_usd": 3000000000,
      "url": "https://www.justice.gov/opa/pr/example",
      "document_title": "Office of Public Affairs press release"
    }
  ],
  "source_coverage": [
    { "source": "courts", "queried": true, "findings_count": 12, "note": null },
    { "source": "sec", "queried": true, "findings_count": 0, "note": "no matching SEC enforcement actions in the current feed window" }
  ],
  "warnings": [],
  "generated_at": "2026-05-29T14:03:11Z"
}

Every field is documented on the Console's Dataset tab. The OUTPUT Markdown report renders the same data as a sourced, readable document.

Example

{
  "entity": "Wells Fargo & Company",
  "sources": ["courts", "doj", "sec", "ftc", "osha"],
  "sinceYear": 2015
}

Via the API:

curl -X POST "https://api.apify.com/v2/acts/shelvick~enforcement-record-profiler/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"entity": "Wells Fargo & Company", "sinceYear": 2015}'

Calling from an AI agent

The Actor is built for agent discovery and invocation.

MCP (mcp.apify.com): surfaces as a callable tool. The input schema is self-documenting — one required field (entity), structured cited output, no follow-up questions — so an LLM can construct correct calls from the tool description alone. Pay per call via the Actor's pay-per-event model, which works with x402 (USDC on Base) and Skyfire managed-token agentic-payment rails.

Apify SDK (Python):

from apify_client import ApifyClient

client = ApifyClient(token=API_TOKEN)
run = client.actor("shelvick/enforcement-record-profiler").call(
    run_input={"entity": "Wells Fargo & Company", "sinceYear": 2015}
)
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["total_findings"], "findings")
    for f in item["findings"]:
        print(f["source"], f["date"], f["title"], "->", f["url"])

REST: /run-sync-get-dataset-items for typical runs; for large entities that query every source, prefer the async /runs endpoint (a full multi-source profile can approach the 5-minute sync cap).

Pricing

Pay-per-event. One charge fires per successful run, after the cited profile is pushed — it covers entity resolution, every selected source query, and the report synthesis. Failed runs are never charged, and the platform startup charge is effectively zero. Subscriber tiers receive a discount.

See the Pricing tab on this Store page for the current per-event rate and any active subscriber discounts.

Behavior

Run-level failures (rare) — only input validation fails a run:

• entity missing or shorter than two characters.
• sources contains a value outside courts / doj / sec / ftc / osha / nlrb / epa.

Per-source degradation (expected) — a source that is unreachable, rate-limited, or returns nothing never fails the run. It is recorded in source_coverage with a note, and an entry is added to warnings. The profile is built from whatever sources did respond. Findings are never invented to fill a gap.

Performance — runtime is set by the slowest selected source. Most return in seconds, and a full multi-source profile typically completes in well under a minute; a single-source run (e.g. ["courts"]) is nearly instant. The NLRB and EPA records are served from indexes refreshed on a schedule, so they answer fast at query time rather than waiting on the source agencies — the trade-off is freshness (a filing from the last few days may not be indexed yet). A very large entity queried across every source can still approach the 5-minute sync-API cap; use the async endpoint when batching.

FAQ

Which sources are covered? Federal courts (CourtListener/RECAP), DOJ, SEC (recent enforcement window), FTC, OSHA, NLRB, and EPA (federal environmental enforcement).

Does it profile individual people? No. Scope is strictly organizations and public figures acting in an official capacity — not private-individual background checks.

Can the summary invent findings? No. Findings are extracted only from the official agency and court pages the Actor fetches, and every finding carries the source URL. The summary is written from those cited findings. Always verify against the linked sources before relying on a finding.

Does it scrape PACER or state courts? No. Court coverage is federal, via CourtListener/RECAP. Direct PACER and state-court portals are out of scope.

Do failed runs cost anything? No. The per-profile charge fires only after a successful run pushes its record; failed runs skip it.

Can I profile several companies at once? No — one entity per run. Run several and combine the dataset records yourself.

What this doesn't do

• No private-individual investigation. Organizations and public figures in an official capacity only.
• EPA coverage is federal environmental enforcement. Civil, administrative, and judicial cases from EPA's federal enforcement docket (Clean Air/Water Act, RCRA, CERCLA, TSCA, etc.), refreshed weekly. State environmental agencies and real-time updates are out of scope.
• No PACER or state-court scraping. Federal dockets come through CourtListener/RECAP; state courts are not covered.
• No legal advice or guilt determination. The Actor reports what the public record states and links to it; interpretation is yours.
• No continuous monitoring. Each run is a one-shot profile, not a standing alert.

For individual background checks, use a person-records service (this Actor deliberately does not). For full regulatory compliance or sanctions screening, use a dedicated compliance platform. For ongoing change alerts on an entity, pair this with a scheduling/monitoring workflow that re-runs it and diffs the results.

---

Design notes: www.scotthelvick.com/tools/enforcement-record-profiler