telmus 0.2.0

Financial statement analysis for AI IDEs and coding agents

telmus

Financial statement analysis for AI IDEs and coding agents.

pip install telmus
telmus scan INFY

What is telmus?

telmus is a Python package and CLI for parsing financial statements, computing key valuation and health ratios, and exposing them through an MCP server for AI tools.

Engines

Engine	What it measures	Key metric
Valuation	Price multiples and peer comparison	P/E, P/B, EV/EBITDA
Health	Balance sheet strength and bankruptcy risk	Piotroski F-score, Altman Z-score
Flags	Earnings quality and cash flow risk	Beneish M-score, free cash flow trend

analyst_brief

A deterministic summary field that explains fundamentals, growth, and red flags without using an LLM.

Example output:

{
  "analyst_brief": "Strong fundamentals (Piotroski F-score of 7). Financially safe (Altman Z-score of 4.20). Revenue growth is 11.2% over three years and operating margins are stable. No significant red flags detected. Suitable for DCF or comparable company analysis."
}

Quick start

1. Install:

pip install telmus

2. Scan a ticker:

telmus scan INFY

3. Read the summary and analyst brief, or export JSON for automation.

MCP server setup

{
  "mcpServers": {
    "telmus": {
      "command": "telmus",
      "args": ["serve"],
      "description": "Financial statement analysis — real ratios for any ticker"
    }
  }
}

• Claude Desktop: use the config to register telmus as an MCP tool.
• Cursor: same config loads telmus as a tool for market research.
• Windsurf: connect the MCP server to get real financial metrics.

Other commands

Command	Description
telmus scan TICKER	Run a full financial scan
telmus scan TICKER --json	Print raw JSON
telmus scan TICKER --export FILE.json	Save JSON to a file
telmus compare A B	Compare two tickers
telmus screen	Run a simple sector screener
telmus serve	Start the MCP server
telmus info	Print package info
telmus check TICKER	Quick health check

Benchmark

Test	Result
Piotroski score coverage	9 signals
Altman Z distress detection	safe/grey/distress ranges
Flag detection	Beneish M, D/E, FCF trend

Architecture

yfinance -> loader -> valuation/health/growth/flags -> ScanResult -> CLI / MCP / SDK

Contributing

1. Fork the repository.
2. Create a feature branch.
3. Run pip install -e ."[dev]".
4. Write tests and update docs.
5. Submit a pull request.

License

MIT License