US Public Company Filing Risk Alerts

Monitor US public-company SEC filing metadata for eight predefined risk categories and return only newly detected risk events after a baseline. Designed for analysts, investor-relations teams, compliance teams, and API/agent workflows that need a small, repeatable filing-change feed.

The Actor reads SEC submissions metadata only. It does not retrieve filing bodies, scrape people, send notifications, or provide investment or legal advice. The first live run establishes a baseline; unchanged later runs produce no Dataset records and no PPE events.

Safe dry-run and Store health check

Set dryRun to true to validate the Actor without making any SEC request, using any SEC contact, or emitting a PPE event. The Actor writes one clearly labeled recordType: "dry-run" fixture so automated Store health checks can confirm a successful non-empty Dataset. This fixture is not SEC data and is not a paid result.

For a live run, set dryRun to false and provide your own SEC requester contact. The contact is used only in the SEC User-Agent header and is never written to the Dataset, KVS, or logs.

What is monitored

The Actor checks Form 8-K item metadata, NT filing-delay notices, and S-1 metadata from the SEC submissions API. The output includes the company, form, filing date, accession number, filing URL, risk category, score, confidence, evidence description, and whether the record is new.

Eight deterministic metadata categories are supported:

Key	Name	Detection condition	Form	Confidence	Evidence	Example false positive
bankruptcy_or_receivership	Bankruptcy or receivership	8-K Item 1.03 appears	8-K	high	Form/item/accession/date metadata	Disclosure may describe a subsidiary or resolved proceeding
debt_acceleration_or_default	Debt acceleration or default	8-K Item 2.04 appears	8-K	high	Form/item/accession/date metadata	Technical default may be cured or immaterial to the whole group
delisting_or_listing_standard	Delisting or listing standard	8-K Item 3.01 appears	8-K	high	Form/item/accession/date metadata	Temporary listing deficiency may be cured
auditor_change	Auditor change	8-K Item 4.01 appears	8-K	high	Form/item/accession/date metadata	Routine auditor rotation may not indicate misconduct
financial_statement_non_reliance	Financial statement non-reliance	8-K Item 4.02 appears	8-K	high	Form/item/accession/date metadata	Subsequent filing may resolve or narrow the issue
officer_departure	Officer departure	8-K Item 5.02 appears	8-K	high	Form/item/accession/date metadata	Ordinary planned executive transition
late_10k / late_10q	Filing-delay notice	NT 10-K or NT 10-Q appears	NT 10-K / NT 10-Q	high	Form/accession/date metadata	Delay can be administrative and later cured
s1_registration	S-1 registration	S-1 appears	S-1	high	Form/accession/date metadata	Registration statement may not complete or may be withdrawn

The categories are screening signals, not conclusions. The Actor does not read filing text, so it cannot determine materiality, intent, outcome, or legal significance.

Pricing

Pay per event is used after monetization is configured:

• company-checked: $0.009 for one successfully saved baseline company-check record.
• change-detected: $0.15 for one newly detected risk-event record after baseline.
• An unchanged run produces 0 chargeable records.
• Failed SEC requests, malformed responses, and skipped inputs produce no chargeable Dataset record.
• The run summary is kept in the named key-value store and is not a chargeable Dataset item.
• Platform usage and named-store retention treatment follow the pricing shown in the Store page. Set a per-run spending limit before running large watchlists.

The two custom PPE events are emitted through the Apify SDK only after the corresponding Dataset record is saved and the returned chargedCount is confirmed. A deterministic record key and named-store ledger prevent repeat charges for the same baseline digest or accession-number risk event.

Input

Use CIKs, tickers, or both. The maximum is 500 distinct companies per run.

{
  "secTickers": ["AAPL", "MSFT"],
  "watchlistId": "my-watchlist",
  "maxCompanies": 500,
  "lookbackDays": 365,
  "maxHistoricalRecords": 1000,
  "firstRunBehavior": "baseline-only",
  "dryRun": false,
  "notificationMode": "disabled",
  "secUserAgentContact": "Your Name your-email@example.com"
}

For live runs, secUserAgentContact must be your own SEC requester contact. It is sent only as the SEC User-Agent header and is not written to the Dataset, KVS, or logs. Never use a fictional contact or another person's contact.

Use a new watchlistId to create a separate baseline. State reset is intentionally disabled because deleting state can create duplicate records and charges.

Output example

{
  "recordKey": "risk:my-watchlist:0000320193:0000320193-26-000001:2.04",
  "recordType": "risk-event",
  "company": "Apple Inc.",
  "cik": "0000320193",
  "ticker": "AAPL",
  "form": "8-K",
  "filing_date": "2026-08-04",
  "accession_number": "0000320193-26-000001",
  "filing_url": "https://www.sec.gov/Archives/edgar/data/320193/000032019326000001/000032019326000001-index.html",
  "risk_category": "debt_acceleration_or_default",
  "risk_score": 85,
  "confidence": "high",
  "evidence": "SEC submissions metadata matched 8-K Item 2.04; filing body was not retrieved.",
  "detected_at": "2026-08-05T00:00:00.000Z",
  "is_historical": false,
  "is_new": true,
  "state_version": 2,
  "chargeable_event": "change-detected"
}

The first run returns a company-check baseline record. Later runs return one risk-event record for each new accession-number/item event. An unchanged run returns no Dataset records; its run summary is stored under run-summary-<runId> in registry-monitor-<watchlistId>.

API and Schedule

Start the Actor with the Apify API. Replace the placeholder contact with your own value and keep it out of public scripts:

curl -X POST "https://api.apify.com/v2/acts/invaluable_rondeau~us-public-company-filing-risk-alerts/runs?token=$APIFY_TOKEN" \
  -H 'Content-Type: application/json' \
  -d '{"secTickers":["AAPL"],"watchlistId":"daily-aapl","secUserAgentContact":"Your Name your-email@example.com"}'

For recurring monitoring, create an Apify Schedule, select this Actor, paste the same input, and start with once daily. A schedule is preferred to a resident worker. Review the first baseline output before increasing the watchlist.

Apify webhooks can be attached by the account owner to a completed or failed run. This Actor does not create or call a webhook and does not send external notifications itself. A webhook recipient should be user-owned and reviewed before activation.

The structured Dataset and run summary can be consumed by an API client, MCP tool, or AI agent. Agents should treat risk records as screening signals and link to the SEC source rather than presenting them as investment conclusions.

SEC access conditions

The SEC documents its EDGAR APIs at EDGAR Application Programming Interfaces. Requests must declare a descriptive User-Agent with contact information and should stay within the SEC's fair-access guidance of no more than 10 requests per second; this Actor uses a minimum 125 ms interval and low concurrency. It retries 429 and 5xx responses with bounded backoff, fails fast on 403, and never bypasses access controls.

The Actor uses the official submissions endpoint and company_tickers.json. It does not retrieve filing bodies. Public SEC information may be copied or further distributed subject to the SEC's published conditions; cite the SEC as the source. This unofficial Actor is not affiliated with or endorsed by the SEC or EDGAR. SEC and EDGAR marks are used only for textual source identification.

Data retention and failure behavior

• State is stored in registry-monitor-<watchlistId>.
• Each chargeable record has a deterministic ledger entry keyed by watchlist, CIK, accession number, item, or baseline digest.
• Dataset push completes before the separate PPE charge call. State advances only after the charge result confirms at least one event, and state-save failures can be retried through the ledger.
• A failed company is placed in the run summary's skipped list; other companies continue.
• A malformed state record stops that company safely instead of overwriting it.
• Inputs above the 500-company limit are rejected or truncated according to the explicit maxCompanies setting; no silent unlimited run is performed.

Limitations

This is metadata monitoring, not a guarantee of financial distress, fraud, default, delisting, or executive misconduct. It may miss signals outside the eight categories, may produce false positives, and does not replace legal, compliance, investment, or accounting review. Do not use it for emergency trading, legal deadlines, or guaranteed alerts.

Changelog

0.2.0

• Fixed state-before-Dataset ordering.
• Added ordered Dataset-then-PPE writes with chargedCount checks.
• Added deterministic ledger and pending-record reconciliation.
• Added bounded retry behavior for 429 and 5xx responses and fail-fast 403 handling.
• Added input/output schemas, eight risk categories, bounded lookback, and 500-company validation.
• Notifications remain disabled.