SEC Filings API
Send a company domain and get back what its 8-K, 10-K, 10-Q and Form D filings say, one finding per record, as structured JSON.
POST/v1/companies/enrich
Replace $AUTOBOUND_API_KEY with your key from the portal.
What's in each record
One record per finding in a filing, with the quoted text and a link to the filing. Each is one of 83 event subtypes, such as acquisitions, restructuring, AI investment and board changes. All 8-K subtypes →
- companies with a recordsince Jan 2026
- 4.3K
- new records per month6-mo avg
- 15K
- have a company domainthe join key
- 99.9%
Companies, records per month and domain rate: audit of all production sec-8k records, Jul 2026.
- signal_subtype
- acquisitionAnnouncedThe event, one of 83 values
- detected_at
- 2026-09-08T07:04:16.976ZWhen Autobound found it (UTC)
- data.summary
- NVIDIA to acquire AI platform Hugging Face for ~$12.9B, including a $1B employee retention program.One sentence that states the finding
- data.sales_relevance
- Integration needs comingThe sales angle in a few words
- data.filing_date
- 2026-09-03Date the filing was made
- data.source_url
- https://www.sec.gov/Archives/edgar/data/1045810/000104581026000078/nvda-20260902.htmThe filing on SEC EDGAR
Also in this record: detail, metrics, excerpts, relevance, sentiment, confidence, filing_year, fiscal_year_end, signal_category, regions_mentioned, vendors_mentioned, technologies_mentioned
signal_id0b57d14a-5b56-4777-b446-6e1b47c2670b
Every field in an SEC filing signal8-K, 10-K and 10-Q share one data shape. Form D has its own.Show all 71 fieldsHide
Green rows are filled in the record selected above: NVIDIA Corporation, 8-K: acquisition.
Signal
signal_idstring, filled in the selected example- Stable id. GET /v1/signals/{id} returns the signal
signal_typestring, filled in the selected example- "sec-8k", "sec-10k", "sec-10q" or "sec-form-d-funding"
signal_subtypestring, filled in the selected example- The event within the type
signal_namestring- Display name. Often empty: use data.summary
detected_atdatetime, filled in the selected example- When Autobound found it (UTC)
associationstring, filled in the selected example- "company" or "contact"
8-K, 10-K and 10-Q fields
data.summarystring, filled in the selected example- One sentence that states the finding
data.detailstring, filled in the selected example- Why it matters to a seller
data.signal_categorystring, filled in the selected example- Theme, such as strategic or workforce
data.excerptsstring, filled in the selected example- The quoted text from the filing
data.metricsobject, filled in the selected example- Numbers the text states: pct, dollar_millions, timeframe
data.filing_datedate, filled in the selected example- Date the filing was made
data.filing_yearinteger, filled in the selected example- Year of the filing
data.fiscal_year_endstring, filled in the selected example- Fiscal year end, MM/DD
data.source_urlstring, filled in the selected example- The filing on SEC EDGAR
data.sentimentstring, filled in the selected example- positive, neutral or negative
data.sales_relevancestring, filled in the selected example- The sales angle in a few words
data.relevancenumber, filled in the selected example- Score from 0 to 1
data.confidencestring, filled in the selected example- Confidence label, such as high
data.vendors_mentionedstring[], filled in the selected example- Vendors named in the text
data.regions_mentionedstring[], filled in the selected example- Regions named in the text
data.technologies_mentionedstring[], filled in the selected example- Technologies named in the text
data.competitors_mentionedstring[]- Competitors named in the text
data.quarterstring- Fiscal quarter of the report (10-Q)
data.fiscal_yearinteger- Fiscal year of the report (10-Q)
data.period_end_datestring- Last day of the period (10-Q)
Form D fields
data.summarystring- Who raised money and how much, in plain words
data.detailstring- Why it matters to a seller
data.total_offering_amountnumber- Offering size in USD, from the filing
data.total_amount_soldnumber- Amount sold so far, in USD
data.total_remainingnumber- Amount still to sell, in USD
data.minimum_investmentnumber- Smallest investment accepted, in USD
data.total_investorsinteger- Investors so far
data.securities_typesstring[]- What was sold, such as Equity
data.exemptionsstring[]- Regulation D exemption codes, such as 06b
data.industry_groupstring- Industry group on the filing
data.entity_namestring- The filer, as named on the filing
data.entity_typestring- Legal form, such as Corporation
data.resolved_entity_namestring- The filer matched to a company
data.resolved_entity_typestring- Company or fund
data.form_typestring- D, or D/A for an amendment
data.file_datedate- Date the form was filed
data.filing_urlstring- The filing on SEC EDGAR
data.accession_numberstring- SEC accession number
data.cikstring- SEC Central Index Key of the filer
data.related_personsobject[]- Officers and directors named on the filing, with roles
data.business_locationstring- City and state of the business
data.business_statestring- State code of the business
data.locationstring- Address city and state, as filed
data.jurisdictionstring- State of incorporation
data.phonestring- Phone number on the filing
data.websitestring- The filer's website
data.year_of_incorporationstring- Year the filer was formed
data.revenue_rangestring- Revenue range on the filing
data.has_non_accredited_investorsboolean- true when some investors are not accredited
data.investment_fund_typestring- Fund type, for a pooled fund
data.aggregate_net_assetsstring- Net asset range, for a fund
data.company_domain_confidencestring- How sure the domain match is
data.sic_codestring- SIC industry code
data.primary_topicstring- Main topic, such as SaaS Funding
data.topicsstring[]- Topic tags
data.relevancenumber- Score from 0 to 1
data.confidencestring- Confidence label, such as high
data.source_urlstring- The filer's page on SEC EDGAR
Company
company.namestring, filled in the selected example- The company the signal is about
company.domainstring, filled in the selected example- Root domain. The join key
company.linkedin_urlstring, filled in the selected example- Company LinkedIn page. May be a relative path
company.industriesstring[]- Industry labels, when known
company.employee_count_*integer- Headcount low and high, when known
company.revenuestring- Revenue range, when known
company.descriptionstring | null- Short description. Often null
Pricing and limits
You pay in credits for what a call returns. A call that returns no results is free.
See all plans →- Enrich a company
- 2creditsper signal returnedPOST /v1/companies/enrich
- Search companies
- 2creditsper company returnedPOST /v1/companies/search
- Free to start
- 1,000free credits at signup
- Self-serve plans
- $19Starter: 1,900 creditsPacks up to $4,999.
- Rate limit
- 300requests per minutePer account, shared by all keys.
Questions
Is there an API for SEC filings by topic?
Yes. The Signal API tags 8-K, 10-K and 10-Q filings with content subtypes, such as aiInvestment, ceoChange or layoffs. Send one companies/search with signal_types ["sec-8k", "sec-10k"] and the subtypes you want. Each result is a company with its matching filings.
Can I get historical SEC filings?
Search looks back 50 days by default. Set detected_after to an earlier date to go further back. The filter uses detected_at, the date Autobound detected the signal, which is not the filing date. Enrich by domain has no default window, so it returns a company's filing signals up to limit.
How is this different from EDGAR?
EDGAR gives you the filings. The Signal API gives you signals from the filings: each one is matched to a company domain and tagged with a subtype and a summary, so you can filter by topic without parsing documents.
What does an SEC filings call cost?
Search costs 2 credits per company returned. Calls with zero results are free, and a new account gets 1,000 free credits.
Make this call with your own key
One key works for all 35 signal types. 2 credits per signal returned with enrich, 2 per company returned with search. Calls with no results are free.
1,000 free credits to start.
Or use it from Claude Code, Cursor or any MCP client
Add the server, run /mcp to sign in, then ask: “What did nvidia.com disclose in its latest 8-K?”
claude mcp add --transport http autobound https://mcp.autobound.ai/mcp