Skip to main content

​
Overview

EP Ratio Screener evaluates stocks on earnings yield and balance-sheet quality. It retrieves fundamentals from Massive, with optional FMP cross-validation, then returns the evaluated universe and the subset that passes the configured thresholds.

​
Connection

MCP URL: https://quantx-api.limex.pro/mcps/ep-ratio-screener Select EP Ratio Screener in the QuantX marketplace. In addition to QuantX gateway authorization, configure these headers in your client’s credentials: The server takes provider keys from the headers of each request, so send them with every call and treat them as secrets: never in tool arguments, webhook query strings, or published JSON. A deployment may also have a key of its own as a fallback, but do not count on one being there. See Client Setup. Cursor configuration (.cursor/mcp.json):
Set the referenced environment variables before starting Cursor. For Claude Code or Python, use the client-specific format in Client Setup.

​
Tools

See Shared MCP Tools for stored results and file export.

​
Examples

​
Check Provider Configuration

Call config-validate with {"request": {}}. This confirms that the required configuration is present, not that every provider API request will succeed.

​
Score One Ticker

Call scoring-engine-score:

​
Screen an Explicit Universe

Call screen-tickers:
An explicit list makes the intended universe clear. To use FMP’s screened universe instead, set use_fmp_screen: true and supply the FMP credential.

​
Results and Limitations

The screen payload includes count, tickers, used_fmp_screen, results, passing_results, passing_tickers, passing_count, flagged_tickers, flagged_passing_tickers, flagged_count, thresholds, and risk_free_rate_pct. A returned result_id must be resolved before formatting the report. Inspect each row for status: "error"; success for some tickers does not mean every provider lookup succeeded. Filter on pass and the returned metrics; do not expect an aggregate score or rank. Keep the flags and the thresholds used with the report, not just the list of passing tickers. The same threshold is reported in two different units, so convert before comparing a config value with a screen row: Row values also carry their scale in the name: market_cap_B is billions and avg_net_income_M is millions. interest_coverage can be a string such as ">50", so check the type before comparing it numerically. To export a CSV, extract passing_results or results and pass that table to save_tool_result as payload, then read the artifact with artifact_read_chunk. Provider coverage, subscriptions, rate limits, and data freshness affect results. Report provider failures separately from stocks that fail the screen. Passing the screen is not a trade instruction.