SEC EDGAR 8-K Material Events Scraper

Turn every SEC Form 8-K into a structured material-event record. Most 8-K feeds stop at the filing metadata and a document URL — this actor opens the filing. You get the reported Item numbers with SEC's own labels, the narrative text under each Item, and the full text of every attached exhibit, including the EX-99 press release that carries the actual earnings numbers and the EX-2/EX-10 agreements that carry the deal terms. Item-code-to-category mapping and boolean flags for the Items that drive alerting (cyber incidents, restatements, bankruptcy, delisting, auditor changes, officer departures) ship as first-class fields, so you can filter on cyber_incident or restatement without maintaining your own Item lookup.

Coverage runs back to EDGAR's modern per-document filing system (~2001) via the quarterly master index, not just a live/recent feed.

What it does

1. Fetches the quarterly master index of every filing for the requested year range and filters to Form 8-K and 8-K/A
2. For each filing, fetches the EDGAR filing-index page for the authoritative Item list (code + SEC's own label), Period of Report, Accepted timestamp, filer SIC/state, and the list of attached documents
3. Fetches the primary document and slices out the narrative text reported under each Item
4. Fetches every exhibit (HTML, legacy .txt, or PDF) and extracts its full text

Input

Field	Type	Default	Description
yearFrom	integer	current year	Start year of the filing range
yearTo	integer	current year	End year (inclusive)
cikList	string list	(blank = all)	Restrict to specific issuer CIKs (e.g. 0000320193 for Apple)
eventCategories	multi-select	(blank = all)	Restrict to filings tagged with at least one of the listed event categories
maxItems	integer	10	Maximum number of 8-K records to return

eventCategories accepts one or more of: earnings, exec_change, m_and_a, bankruptcy, delisting, auditor_change, material_agreement, impairment, cyber_incident, restatement, debt, other.

Output schema

Each record represents one 8-K (or 8-K/A) filing.

Field	Description
accession_number	SEC EDGAR accession number
cik	Issuer CIK, zero-padded to 10 digits
company_name	Issuer name as filed
ticker	Exchange ticker, when resolvable
sic_code / sic_label	SEC industry code and label
state_of_incorporation	Issuer's state or foreign jurisdiction of incorporation
form_type	8-K or 8-K/A
filed_date	Date received by EDGAR
period_of_report	Date of the reported event
acceptance_datetime	EDGAR acceptance timestamp
filing_index_url	URL to the EDGAR filing-index page
primary_document_url	URL to the primary 8-K document
item_codes / item_labels	Reported Item numbers and SEC's own labels, in matching order
event_category	One or more category tags derived from item_codes
item_body_text	Narrative text sliced from the primary document, per reported Item
exhibits	[{ exhibit_type, description, url, text, byte_size, parse_status }]
press_release_text	Extracted text of the first EX-99.x exhibit, when present
is_amendment	true when form_type is 8-K/A
amends_accession	Best-effort accession number of the original 8-K this amendment supersedes (blank when unresolvable — never guessed)
has_financial_statements	true when Item 9.01 was reported
cyber_incident_flag	Item 1.05 — SEC's cyber-disclosure rule
auditor_change_flag	Item 4.01
restatement_flag	Item 4.02
bankruptcy_flag	Item 1.03
delisting_flag	Item 3.01
officer_departure_flag	Item 5.02

Notes

• Coverage: roughly 40,000 8-K/8-K-A filings per year across all reporting issuers. The quarterly master index (form.idx) is a single flat file per quarter — there is no per-issuer or per-category partitioning to work around, so a year range covers everything filed in it.
• Exhibit text: HTML and legacy .txt exhibits return full text. PDF exhibits are extracted locally for born-digital PDFs; a scanned/image-only PDF returns

  parse_status: "pdf_no_extractable_text"
  rather than a fabricated body.
• Rate compliance: the actor stays comfortably under SEC EDGAR's published fair-access request-rate ceiling and sends the required declared identification header on every request — without it, EDGAR returns 403 on every path.
• Historical boundary: item/exhibit extraction depends on EDGAR's per-document filing-index page, a feature of the modern (~2001+) filing system. Pre-2001 accessions enumerated from the master index but lacking a filing-index page are skipped rather than parsed from the legacy compound submission format.
• Amendment linkage (amends_accession) is best-effort: EDGAR carries no formal "amends" pointer on an 8-K/A, so the actor looks up the issuer's recent filing history and takes the most recent prior 8-K filed before the amendment. Left blank when unresolvable.

Use cases

• Event-driven funds: Screen for M&A, restatements, or debt triggers the moment they're filed
• Third-party risk / vendor monitoring: Alert on Item 1.05 material cybersecurity incidents across a vendor or portfolio company list
• M&A trackers: Pull the full text of merger agreements and press releases, not just the filing metadata
• Financial newsrooms: Get the earnings press release text (EX-99.x) as soon as it's filed
• Compliance teams: Track auditor changes, delistings, and bankruptcy disclosures across a watchlist of issuers