mcp

Use when: You want Cursor, Claude, or another MCP client to call Disclosure Alpha tools — either by ticker (analyst) or from raw HTML (builder).

Start here

Analyst bundle (disclosure-alpha-mcp-analyst) — agent knows tickers:

  • list_company_filings_tool — filing index for a ticker

  • score_company_filing_tool — score by ticker + fiscal year

Builder bundle (disclosure-alpha-mcp-builder) — agent has filing HTML:

  • extract_sections_tool, compute_section_metrics_tool_wrapper, diff_sections_tool

  • score_deterministic_tool_wrapper, score_filing_html_tool_wrapper

Configure SEC_USER_AGENT in the MCP server environment. User guide: MCP Guide.

Example

export SEC_USER_AGENT="YourName your@email.com"
disclosure-alpha-mcp-analyst   # ticker discovery + scoring
disclosure-alpha-mcp-builder     # raw HTML pipeline steps

Full API

Analyst bundle

Analyst MCP bundle: ticker discovery + scoring (2 tools + taxonomy).

EDGAR-backed tools support 10-K and 10-Q only. For 8-K, use the Builder bundle with local HTML.

disclosure_alpha.mcp.analyst.list_company_filings_tool(ticker: str, fiscal_year: int, form_type: str | None = None) → str[source]

List available 10-K / 10-Q filings for a ticker and fiscal year.

disclosure_alpha.mcp.analyst.score_company_filing_tool(ticker: str, fiscal_year: int, form_type: str = '10-K', quarter: str | None = None, scoring_model_version: str = 'deterministic_scoring_v2') → str[source]

Score a company filing by ticker and fiscal year (10-K or 10-Q with quarter).

disclosure_alpha.mcp.analyst.taxonomy() → str[source]

Score taxonomy: component weights and version strings.

disclosure_alpha.mcp.analyst.main() → None[source]

Builder bundle

Builder MCP bundle: low-level pipeline tools (5 tools, no ticker helpers).

disclosure_alpha.mcp.builder.extract_sections_tool(html: str, form_type: str) → str[source]

Extract SEC filing sections from HTML (10-K, 10-Q, or 8-K; 8-K: local HTML only).

disclosure_alpha.mcp.builder.compute_section_metrics_tool_wrapper(sections_json: str, prior_sections_json: str | None = None) → str[source]

Compute deterministic text metrics and diffs from extracted section payloads.

disclosure_alpha.mcp.builder.diff_sections_tool(current_text: str, prior_text: str, section_name: str = 'section') → str[source]

Diff two section texts and return change score + language deltas.

disclosure_alpha.mcp.builder.score_deterministic_tool_wrapper(metrics_json: str, scoring_model_version: str = 'deterministic_scoring_v2') → str[source]

Aggregate deterministic component scores from a metrics payload.

disclosure_alpha.mcp.builder.score_filing_html_tool_wrapper(html: str, form_type: str, prior_html: str | None = None, scoring_model_version: str = 'deterministic_scoring_v2') → str[source]

Run full pipeline on filing HTML (10-K, 10-Q, or 8-K; 8-K: local HTML only).

disclosure_alpha.mcp.builder.main() → None[source]

Tool implementations: disclosure_alpha.mcp.tools.