pyBuffet

EDGAR Ledger — a single-file dashboard that pulls company financials (income statement, balance sheet, cash flow) straight from SEC EDGAR's free XBRL APIs, plus a tiny stdlib relay that satisfies the SEC's declared User-Agent requirement.

Quick start

1. Open edgar_relay.py and set IDENTITY to your contact email (the SEC wants to know who is calling).

2. Run the relay from this folder — no packages needed:

   python edgar_relay.py
   

   It serves the dashboard at http://127.0.0.1:8787/ and forwards SEC calls with the proper header.

3. In the dashboard, load a watchlist: drop in your own .json or click Load the example list (the bundled investments.json).

Watchlist format

The dashboard asks for a .json file shaped like investments.json:

{
  "investments": [
    {
      "company": "Micron Technology",
      "category": "RAM",
      "ticker": "MU",
      "exchange": "NASDAQ",
      "country": "United States",
      "date_spotted": "01/02/2026"
    }
  ]
}

All six keys are strings. Use "NONE" where a field doesn't apply (e.g. exchange for indices, commodities, FX or crypto) and MM/DD/YYYY or "NONE" for date_spotted. Entries in the same category should sit next to each other — the sidebar groups them in order.

Building a list with an AI

Open the Build a list (AI) tab in the dashboard. Copy the prompt shown there into any AI assistant (Claude, ChatGPT, Gemini, …), then paste your own list of companies / tickers / notes directly underneath the prompt and send it. Save the AI's JSON reply as a .json file and load it from the Ledger tab.

Warren Buffett tab

The Warren Buffett tab holds chapter notes from Warren Buffett and the Interpretation of Financial Statements (chapters 1–13). Each chapter block shows the chapter number, title, exhibit tables and summary. The source data lives in buffett_book_ch113.json and is embedded in the page so it works offline.

Design

The dashboard uses a boxed-mosaic editorial system: black-on-white two-colour discipline (ink #111111 on paper #FFFFFF, two greys for secondary text), 1px square ink borders on every module, deep burgundy reserved for data highlights, and three type voices — Afacad Flux (uppercase micro-labels/buttons), Newsreader (body prose and italic captions), Prata (display headings). The ◐ Dark/Light button in the header toggles a pure inversion, persisted in localStorage.

Files

File	What it is
edgar-ledger.html	The dashboard — watchlist sidebar, financials ledger, AI prompt tab, Buffett chapter notes
edgar_relay.py	Local relay/server (stdlib only) that adds the SEC User-Agent header
investments.json	Example watchlist in the expected format
buffett_book_ch113.json	Chapter notes shown in the Warren Buffett tab