lamm-mit/skills/edgartools

EDGARTools

Python library for accessing SEC EDGAR filings (1994–present). Company financials, insider trading, institutional holdings, and more without manual EDGAR navigation.

Installation

pip install edgartools

Required Setup

from edgar import set_identity

# REQUIRED: EDGAR requires identification for all requests
set_identity("Your Name [email protected]")

Core Usage

Company Lookup

from edgar import Company, find_company

# By ticker symbol
apple = Company("AAPL")
nvidia = Company("NVDA")

# By CIK number
company = Company(320193)

# Search by name
results = find_company("Moderna")

Filing Retrieval

company = Company("MRNA")

# Get all filings
filings = company.get_filings()

# Filter by form type
annual_reports = company.get_filings(form="10-K")
quarterly_reports = company.get_filings(form="10-Q")
current_reports = company.get_filings(form="8-K")
proxy_statements = company.get_filings(form="DEF 14A")

# Get most recent filing
latest_10k = company.get_filings(form="10-K").latest(1)

# Get filings from date range
filings_2023 = company.get_filings(form="10-K", date="2023-01-01:2023-12-31")

Financial Data (XBRL Parsing)

# Get financial statements from 10-K
tenk = company.get_filings(form="10-K").latest(1)
filing = tenk[0]

# Access XBRL financial data
financials = filing.financials

income_stmt = financials.income_statement
balance_sheet = financials.balance_sheet
cash_flow = financials.cashflow_statement

# As pandas DataFrame
import pandas as pd
df = income_stmt.to_dataframe()
print(df[["NetIncomeLoss", "Revenues", "GrossProfit"]])

# Key metrics
revenue = income_stmt["Revenues"].value
net_income = income_stmt["NetIncomeLoss"].value
total_assets = balance_sheet["Assets"].value

Insider Trading (Form 4)

# Get recent insider transactions
insider_trades = company.get_filings(form="4")

for filing in insider_trades[:10]:
    transaction = filing.obj()
    print(f"Filer: {transaction.reporting_owner}")
    print(f"Type: {transaction.transaction_type}")  # Buy/Sell
    print(f"Shares: {transaction.transaction_shares}")
    print(f"Price: {transaction.transaction_price}")
    print(f"Date: {transaction.transaction_date}")

13F Institutional Holdings

# Get hedge fund / institutional holdings
blackrock = Company("BLK")
form_13f = blackrock.get_filings(form="13F-HR").latest(1)[0]

holdings = form_13f.obj()
holdings_df = holdings.to_dataframe()

# Top holdings
print(holdings_df.nlargest(20, "value")[["nameOfIssuer", "shares", "value"]])

Full-Text Search

from edgar import full_text_search

# Search across all SEC filings
results = full_text_search("mRNA vaccine efficacy", form="10-K", date_range="2023:2024")

Practical Examples

Biotech Drug Pipeline from 10-K

company = Company("MRNA")
filing = company.get_filings(form="10-K").latest(1)[0]

# Access full document text
doc = filing.document
text = doc.text()

# Search for pipeline section
import re
pipeline_section = re.search(r"(Pipeline|Product Candidates)(.*?)(?=\n#{1,3} )",
                              text, re.DOTALL)

Track Insider Selling Before Drug Approval

company = Company("BIIB")  # Biogen
insider_trades = company.get_filings(form="4", date="2023-01-01:2024-01-01")

sells = []
for f in insider_trades:
    t = f.obj()
    if hasattr(t, 'transaction_type') and 'S' in str(t.transaction_type):
        sells.append({
            "owner": t.reporting_owner,
            "shares": t.transaction_shares,
            "date": t.transaction_date
        })

Multi-Year Revenue Trend

company = Company("PFE")  # Pfizer
annual_filings = company.get_filings(form="10-K")[:5]  # Last 5 years

revenues = []
for f in annual_filings:
    fin = f.financials
    revenues.append({
        "year": f.filing_date.year,
        "revenue": fin.income_statement["Revenues"].value
    })

Output Format

All financial data returns as structured Python objects with .to_dataframe() support. Filings return Filing objects with .obj() for parsed content, .text() for raw text, .document for full filing.

Limitations

• Some older filings (pre-2009) lack XBRL structured data
• Rate limits apply (EDGAR enforces ~10 req/sec)
• Must call set_identity() or requests will be blocked