Direct Answer
SEC EDGAR is the free, official system for finding every U.S. public company's filings. Start by resolving the company's exact name and 10-digit Central Index Key (CIK) rather than guessing a URL from its ticker, then choose the right search: EDGAR full-text search looks inside the words of filings, while Company Search lists everything one filer has submitted, in order. For financial figures, SEC's free XBRL APIs (data.sec.gov) return structured, machine-readable data tagged with its own units, fiscal period, and source filing, with no API key required.
Key Takeaways
- EDGAR is the SEC's official filing system, not a paid data vendor - every 10-K, 10-Q, 8-K, and other filing is submitted through it and published there first.
- EDGAR full-text search and Company Search solve different problems: full-text search indexes the words inside filings, while Company Search indexes only filing header metadata for one specific filer.
- A company's CIK (Central Index Key), not its ticker, is the durable identifier EDGAR and its APIs use - tickers change, get reused, or can apply to more than one security.
- Every filing has an accession number built from the filer's CIK, the filing year, and a sequential count. Recording it, not just "the 10-K," is what makes a research note reproducible.
- SEC's structured XBRL data (the submissions, companyfacts, and companyconcept APIs) is free and requires no API key, but automated access must declare a descriptive User-Agent and stay under 10 requests per second.
What Is EDGAR, and What Can You Actually Get From It?
Direct answer: EDGAR (Electronic Data Gathering, Analysis, and Retrieval system) is the SEC's official electronic filing system, operating since the mid-1990s. Every U.S. public company, fund, and individual with a disclosure obligation files through it, and the SEC publishes the resulting filings on EDGAR as the source of record. Government-created content on sec.gov and EDGAR public filing content are free to access and reuse.
EDGAR provides two different layers of information, and confusing them is the fastest way to waste research time. The first layer is the filings themselves - the actual 10-K, 8-K, or proxy statement document, in HTML or PDF, exactly as filed. The second layer is structured data about those filings: metadata such as filing dates and form types, and extracted XBRL financial facts pulled out of the statements themselves and made available as JSON through SEC's public APIs. A company's own investor-relations page often mirrors the same document, but EDGAR remains the authoritative source - the filing timestamp, accession number, and exhibit list on EDGAR are the version of record.
Finding the Right Company: Ticker, CIK, and Company Search
Two EDGAR entry points cover most company-lookup needs, and they answer different questions:
- Company Search (sec.gov/cgi-bin/browse-edgar?action=getcompany) - enter a company name or ticker to reach that filer's EDGAR landing page, listing every filing it has made, sorted by date and filterable by form type. This is the starting point for "show me everything this company has filed."
- EDGAR full-text search (sec.gov/edgar/search/) - searches the actual text inside filings, across one company or many. This is the starting point for "find where this specific language appears," covered in more detail below.
Behind both searches sits the CIK: a 10-digit Central Index Key EDGAR assigns to each filer (padded with leading zeros in API URLs, for example 0000909494). A CIK is the durable identifier - it does not change when a company changes its ticker, gets relisted after a delisting, or has its old ticker reassigned to an unrelated company. SEC also publishes bulk ticker-to-CIK mapping files (ticker.txt and company_tickers.json) for programmatic lookups, but the SEC's own documentation notes these are periodically updated without a guarantee of accuracy or scope - treat them as a starting point, not the final word, for a lookup that matters.
Similarly named companies, subsidiaries that file separately from a parent, and holding-company restructurings can all make the "obvious" CIK the wrong one. Confirm the CIK from the filer's own EDGAR landing page - checking the company name, former names, and recent filing history shown there - before recording it as evidence in a research note.
How EDGAR Full-Text Search Works
EDGAR full-text search (sec.gov/edgar/search/) indexes the actual text content of filings - useful for finding a specific disclosure, phrase, or risk factor across one company or across many, rather than browsing one filer's history. This is a meaningfully different tool from Company Search, which indexes only filing header metadata, not the words inside the document itself.
The search supports Boolean operators in all capital letters (AND, OR, NOT), exact-phrase matching with quotation marks, and filtering by form type, date range, and entity. Common short words ("the," "is," "at," "which," "on") are treated as stop words and are not indexed, so a phrase search built around them can miss matches. The index also reaches into XML-tagged content inside a filing - for example, matching "hedge fund" inside a Form D's structured fund-type fields - though it searches the tagged text itself, not the underlying XML element names.
Full-text search does not cover the earliest years of EDGAR's history. For a company's complete filing history, including older filings that predate full-text search's coverage, Company Search or the daily/full index files are the more reliable path.
Reading the Filing Index: Accession Numbers and the Filing Detail Page
Every accepted EDGAR submission gets an accession number - a unique identifier such as 0001193125-15-118890, and its structure is documented, not arbitrary. The first 10 digits are the CIK of the entity that submitted the filing (which can be a third-party filer agent rather than the company itself). The next 2 digits are the year. The remaining digits are a sequential count of that CIK's submitted filings, which usually - but not always - resets to zero at the start of each calendar year.
Any individual filing document's URL can be turned into its filing detail page (the index listing every document and exhibit in that submission, plus a link to the filer's complete history) by reformatting the accession number segment of the path: insert dashes after the 10th and after the 12th digit, and append "-index.htm". For example, an exhibit at .../data/909494/000143774919004069/ex_136497.htm has its filing detail page at .../data/909494/0001437749-19-004069-index.htm.
Directory browsing is allowed for the CIK-specific subdirectories under /Archives/edgar/data/ (the individual filer folders), though the top-level /data/ directory itself is not browsable given the volume of CIK subdirectories it would contain. Recording the exact form type, filing date, and accession number for any filing cited in research is the same discipline the site's SEC Filing Comparison Workflow guide recommends when comparing two versions of a filing - a description like "the recent 10-K" is not reproducible; an accession number is.
XBRL and the Structured-Data APIs
XBRL (eXtensible Business Reporting Language) is the XML-based tagging standard the SEC has required for financial-statement reporting since 2009, now typically embedded as inline XBRL directly inside the HTML filing rather than a separate file. Each reported fact is tagged against a standard taxonomy - us-gaap, ifrs-full, dei, or srt - or, in some cases, against a company's own custom taxonomy extension. Facts using a standard taxonomy are comparable across companies and time in a way custom-tagged facts are not guaranteed to be, which is exactly why SEC's cross-company APIs described below only aggregate the standard-taxonomy facts.
SEC's data.sec.gov APIs return this data as JSON, require no authentication or API key, and are updated in near real time as filings are processed:
| Endpoint | Returns |
|---|---|
| data.sec.gov/submissions/CIK##########.json | A filer's current name, former names, exchanges, tickers, and recent filing history in one call. |
| data.sec.gov/api/xbrl/companyfacts/CIK##########.json | Every standard-taxonomy XBRL fact a company has ever reported, in a single call. |
| data.sec.gov/api/xbrl/companyconcept/CIK##########/us-gaap/{Tag}.json | One specific tag's full reporting history for one company, for example AccountsPayableCurrent. |
| data.sec.gov/api/xbrl/frames/us-gaap/{Tag}/{Unit}/CY####Q#I.json | One fact per reporting entity that best fits a given calendar period, across all filers - useful for a single-period, cross-company comparison rather than one company's own history. |
data.sec.gov does not support Cross-Origin Resource Sharing (CORS), so these APIs cannot be called directly from a page's client-side JavaScript running in a browser - any tool built on this data has to fetch and cache it server-side. For bulk access to large amounts of data, SEC republishes companyfacts.zip and submissions.zip nightly (approximately 3:00 a.m. ET) as an alternative to many individual live API calls.
Rate Limits, User-Agent Requirements, and Automated Access
SEC's published fair-access guidelines limit each user to no more than 10 requests per second in total, regardless of how many machines are used to submit requests, in order to preserve equitable access to EDGAR for everyone. The SEC reserves the right to block IP addresses that submit excessive requests, and does not allow undeclared or "unclassified" automated tools to crawl the site.
Automated requests are expected to declare a descriptive User-Agent header identifying the requester - SEC's own documentation gives a sample format of a company name and contact email, for example Sample Company Name AdminContact@sample-company-domain.com. A request without a declared user agent risks being rejected with an "Undeclared Automated Tool" error. No API key or other authentication is required, only this identification and a moderated request rate. SEC explicitly does not provide technical support for building or debugging scripted downloads, which makes getting this right the requester's own responsibility.
Common Mistakes and How to Avoid Them
| Mistake | Why it causes problems | Better practice |
|---|---|---|
| Guessing a company's EDGAR page from its ticker | Tickers can be reused after a delisting, apply to more than one class of stock, or belong to a subsidiary that files separately from its parent. | Resolve and record the company's CIK through Company Search first, and confirm it from the filer's own landing page. |
| Treating full-text search as a complete filing history tool | Full-text search indexes filing text but does not reach back through EDGAR's earliest years, and it is a different index from a filer's own chronological filing list. | Use Company Search or the daily/full index files for a company's complete filing history; use full-text search to find specific language. |
| Assuming any XBRL-tagged number is comparable across companies | SEC's cross-company APIs (companyfacts, frames) only aggregate facts tagged with a standard taxonomy precisely because custom-tagged extensions are not guaranteed comparable. | Check whether a reported figure uses a standard us-gaap/ifrs tag or a company-specific extension before comparing it across companies. |
| Scripting requests without a declared User-Agent or rate limit | Undeclared automated tools can be blocked outright, and SEC offers no technical support for debugging a script that runs afoul of this. | Send a descriptive User-Agent identifying the requester and stay within the 10-requests-per-second guideline. |
| Citing "the 10-K" without an accession number | A company can file amendments (10-K/A) that share a similar title but represent a materially different version. | Record the exact accession number and filing date alongside the form type, not just the form type and year. |
Risks and Limitations
A filing's acceptance by the EDGAR Filer System and its actual availability on sec.gov are not simultaneous - SEC's own documentation describes a typical lag of one to three minutes, which can grow significantly during periods of high server load. A research process that depends on being first to a filing should not assume instant availability.
The bulk ticker-to-CIK mapping files SEC publishes are periodically updated without a guarantee of accuracy or scope, so a lookup that matters for an unfamiliar or thinly traded company deserves a manual confirmation through Company Search, not sole reliance on a bulk file. Similarly, the structured XBRL APIs return exactly what a company tagged and reported - they do not interpret, verify, or flag whether a number tells the whole story, which is why footnotes, MD&A, and the filing text itself remain necessary reading rather than something these APIs can substitute for.
None of this is investment advice or a recommendation. EDGAR access is a research skill, not a signal in itself - what a filing or a data point means for a specific decision still requires the analysis this cluster's other guides cover.
Frequently Asked Questions
What is SEC EDGAR?
EDGAR (Electronic Data Gathering, Analysis, and Retrieval system) is the SEC's free, official system for collecting and publishing the filings public companies, funds, and individuals are required to submit. It has operated since the mid-1990s, and government-created content on sec.gov and EDGAR filings are free to access and reuse. EDGAR is the primary source of record; third-party aggregators copy or reformat the same underlying filings.
What is a CIK, and why does it matter more than a ticker symbol?
A CIK (Central Index Key) is the 10-digit identifier EDGAR assigns to each filer, and it is the durable identifier EDGAR's searches and APIs are built around. A ticker symbol can change, get reused by a different company after a delisting, or apply to more than one class of stock, but a CIK does not - resolving and recording the correct CIK first makes a research trail reproducible in a way a ticker alone does not.
Do I need an API key to use SEC EDGAR's data APIs?
No. The data.sec.gov submissions and XBRL APIs (companyfacts, companyconcept, and frames) return JSON with no authentication or API key required. Automated requests must still declare a descriptive User-Agent header identifying the requester and stay within the SEC's published guideline of no more than 10 requests per second - requests without a declared user agent can be blocked as an undeclared automated tool.
What does an EDGAR accession number tell you?
An accession number such as 0001193125-24-118890 has three parts: the first 10 digits are the CIK of the entity that submitted the filing (which can be a third-party filer agent rather than the company itself), the next 2 digits are the year, and the remaining digits are a sequential count of that CIK's filings, which usually but not always resets to zero at the start of each calendar year. Recording the full accession number, not just the form type and year, is what makes a filing reference precise and reproducible.
How does full-text search differ from browsing a company's filing list?
Full-text search locates specific language across companies and filings, which is how you find every filing mentioning a particular term, contract type, or counterparty. Browsing a company's list shows what that company filed and when. The two answer different questions, and full-text search is the less used and often more powerful of the two.
What structured data is available beyond the documents themselves?
Financial statement data is available in structured form derived from the tagged filings, which allows figures to be extracted programmatically rather than read from documents. This supports building time series across companies without manual entry. The structured data reflects what companies tagged, so it inherits any tagging inconsistencies.
How far back does the filing archive extend?
Electronic filing was phased in during the mid-1990s, so coverage is essentially complete from that period forward and sparse before it. For older history, other sources are required. This matters when examining a company across a full cycle, since a cycle beginning before electronic filing may be partially undocumented in the archive.
What identifier should be used when tracking a company over time?
The central index key assigned to each filer, because it persists through name changes, ticker changes, and reincorporations, while tickers are reassigned and names change. Building any tracking system on the ticker rather than the persistent identifier eventually produces broken links. The identifier appears on every filing.
How quickly do filings become publicly available after submission?
Filings are generally disseminated shortly after acceptance, with acceptance times recorded on each filing. Documents submitted outside business hours may be dated the following business day. For time-sensitive filings such as material event reports, the acceptance timestamp indicates when the information actually became public.