Disclosure Alpha¶

Parse SEC filing HTML, compute deterministic text metrics, diff sections, and produce reproducible disclosure risk scores — no LLM required.

Scope

Not investment advice. Scores summarize language and change signals in filings. Full claims and limits: What This Does and Does Not Claim.

Start here¶

I want to…	Start here
Prove it works in five minutes	First Successful Run
Evaluate whether to trust this	Evidence and Validation
Understand the numbers	Understanding Scores
Score in terminal	Quickstart: CLI
Build a screener	HTTP API Guide → Workflows
Use in Python	Quickstart: Python API
Wire an agent	MCP Guide
Copy-paste examples	Examples Gallery

1. Install (Python 3.11+)

pip install "disclosure-alpha"
export SEC_USER_AGENT="YourName your@email.com"

2. Score a filing

disclosure-alpha score --ticker AAPL --fiscal-year 2025 --form 10-K

Or score local HTML with no network: First Successful Run.

3. Pick your integration surface

CLI, Python SDK, HTTP API, or MCP — see Choose Your Surface.

Full install options: Installation.

Deterministic scores — ten computed components (nine headline-weighted, 0–100) plus an overall disclosure risk score; see Score Catalog
Section extraction — Item 1A, MD&A, controls, and more from 10-K / 10-Q HTML (8-K via local --html or MCP Builder only; not EDGAR or HTTP ticker routes)
Change detection — lexical and semantic diffs vs the prior comparable filing
Multiple surfaces — terminal CLI, Python imports, REST API, and MCP tools for agents