Thesma MCP Server

Give your AI assistant access to SEC, Census, BLS, and SBA data.

What it does

An MCP server that wraps the Thesma API, giving AI assistants (Claude, Cursor, ChatGPT) native access to SEC EDGAR filings — US-GAAP 10-K and IFRS 20-F, covering every US-listed public company on NYSE and NASDAQ (~6,000 companies) — plus Bureau of Labor Statistics employment data, US Census Bureau demographics, and SBA 7(a) small business lending. Ask questions in plain English, get structured data back.

Quick example

"What was Apple's revenue last year?"

The AI calls get_financials and returns Apple's income statement with formatted line items.

"What was Spotify's revenue last year?"

The AI calls get_financials for SPOT and returns Spotify's income statement in EUR with taxonomy: ifrs-full and by_nature presentation metadata.

"Find high-margin S&P 500 companies where insiders are buying"

The AI calls screen_companies with margin filters and insider buying signals.

"Which funds increased their position in NVDA last quarter?"

The AI calls get_holding_changes and shows quarter-over-quarter position changes.

"What's the average wage for software developers in Texas?"

The AI calls get_occupation_wages with SOC code 15-1252 and state filter, returning median and percentile wage data.

"What does Apple's labor market look like — hiring trends, local wages, compensation benchmarks?"

The AI calls get_company for AAPL, which automatically includes BLS labor market context alongside SEC company details.

"Give me a full picture of Apple — profile, financials, ratios, insider trades, top holders, executive compensation, and board."

The AI calls get_company('AAPL', include='financials,ratios,insider_trades,holders,compensation,board,labor_context,lending_context') and returns a composed view in one call instead of eight tool invocations.

Installation

pip install thesma-mcp

Claude Desktop

Add to your config file (~/Library/Application Support/Claude/claude_desktop_config.json on macOS, %APPDATA%\Claude\claude_desktop_config.json on Windows):

{
  "mcpServers": {
    "thesma": {
      "command": "uvx",
      "args": ["thesma-mcp"],
      "env": {
        "THESMA_API_KEY": "your-api-key"
      }
    }
  }
}

Cursor

Add to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "thesma": {
      "command": "uvx",
      "args": ["thesma-mcp"],
      "env": {
        "THESMA_API_KEY": "your-api-key"
      }
    }
  }
}

Using pip install instead of uvx? If you've already installed thesma-mcp with pip, you can use "command": "thesma-mcp" directly (no args needed) instead of uvx.

Get your API key at portal.thesma.dev (free tier: 250 requests/day).

Available tools

Company Discovery

Financial Statements

Financial Ratios

Screening

Corporate Events

Insider Trading

Institutional Holdings

Compensation & Governance

Filings

Filing Section Search

Webhooks

Industry Lookup

Industry Employment (CES)

County Employment & Wages (QCEW)

Occupation Wages (OEWS)

Labor Market Turnover (JOLTS)

Local Area Unemployment (LAUS)

SBA 7(a) Lending

BLS Discovery

Configuration

Data coverage

• ~6,000 US-listed public companies on NYSE and NASDAQ — US-GAAP filers plus IFRS 20-F filers (Spotify, Nu Holdings, ASML, and other US-listed IFRS reporters). Native-currency reporting; no USD normalization.
• SEC EDGAR: financial statements (2009-present), insider trades, institutional holdings, executive compensation, board data, corporate events, filings
• Bureau of Labor Statistics: industry employment (CES), county wages (QCEW), occupation wages (OEWS), job openings and turnover (JOLTS), local unemployment (LAUS)
• Small Business Administration: 7(a) loan program data — quarterly lending aggregates by county/state/industry, lender rankings, loan characteristics (size, term, rate, sub-programme), vintage cohort charge-off outcomes
• Labor & lending market enrichment: get_company automatically includes BLS labor context and SBA 7(a) lending context; screen_companies supports labor market and SBA lending filters
• Known gaps: IFRS bank filings (Nu Holdings, HSBC, Deutsche Bank, etc.) — the core banking-statement extensions (net_interest_income, etc.) are US-GAAP-specific today; IFRS banking fields land in a future release.
• All data sourced from US federal public-domain sources: SEC EDGAR, US Census Bureau, Bureau of Labor Statistics, Small Business Administration

Links

• Thesma API docs
• Developer portal
• Pricing
• Security & data rights
• Website

License

MIT