Parallax ExtremeHurst Server

Overview

The Parallax ExtremeHurst Server generates ExtremeHurst signals from OHLCV data. It can run directly on prepared bars or pull data through the Dataloader Server before running the local signal engine.

Connection

Add this server to your MCP client configuration.

Cursor
Bearer auth

{
  "mcpServers": {
    "parallax-extreme-hurst": {
      "url": "coming soon"
    }
  }
}

{
  "mcpServers": {
    "parallax-extreme-hurst": {
      "url": "coming soon",
      "headers": {
        "Authorization": "Bearer ${env:PARALLAX_EXTREME_HURST_MCP_TOKEN}"
      }
    }
  }
}

Restart the client after changing MCP configuration. The server tools appear automatically after the connection is established.

Transport

Property	Value
Protocol	MCP over Streamable HTTP
MCP URL	`coming soon`
Health URL	`https://parallax-hurst-signals-production.up.railway.app/health`
MCP path	`/mcp/parallax-extreme-hurst-mcp-server`
Request envelope	`{"request": {...}}`
Auth	Optional bearer token when enabled for the endpoint

Best-Fit Workflows

Generate ExtremeHurst signals from ticker/date-range input.
Analyze a shared-workspace Parallax TXT file.
Run the signal engine on prepared bar arrays.
Inspect or download large signal payloads with shared result/artifact tools.

Recommended Tools

pull_and_generate_signals
generate_signals_from_workspace_file
generate_signals
deployment_info
runtime_health
ping

Tools

Tool	Description	Returns
`ping`	Health check endpoint that returns the server’s current UTC time. Params: none Returns: Any	`Any`
`generate_signals`	POST /api/v1/Signals/generate - Run the ExtremeHurst engine to produce trading signals from OHLCV data. Returns signal metadata, processing stats, and CRC validation. Params: payload: Any - JSON request body containing OHLCV bars and config; example: {“security”: “AAPL”, “interval”: “DAILY”, “bars”: […]} Returns: Any	`Any`
`generate_signals_from_workspace_file`	Read a Parallax TXT OHLCV file from the scoped workspace and generate signals using the local ExtremeHurst backend. Use this for workspace-relative files created via fs_write_file or already present in shared storage. Params: path, security?, quality?, preferences?, last_signal_only?, history_bars?.	`dict[str, Any]`
`pull_and_generate_signals`	Call dataloader-mcp-server fetch-ohlcvs directly, convert the returned rows into OHLCV arrays, and generate Parallax signals locally. This returns signals and provenance, not a raw-data export file. If the user wants a local CSV/Parquet download, use dataloader-mcp-server save-dataframe instead of manually writing rows. Params: ticker, start_date, end_date, source?, quality?, preferences?, last_signal_only?, history_bars?, refresh?. `refresh` is retained for compatibility and is not currently forwarded to the dataloader.	`dict[str, Any]`
`load_bars_from_txt`	Parses a server-local .txt OHLCV data file into separate arrays of timestamps, opens, highs, lows, closes, and volumes. This reads filesystem paths visible to the Parallax server process, not workspace-relative files created via fs_write_file. Params: path: file path to the .txt data file (e.g., ‘data/samples/sample.txt’) Returns: Any	`Any`
`print_signal`	Prints a formatted SignalSummary object showing trading signal details including name, direction, rank, edge, price, and timing information. Params: i (int): Signal index; sig (SignalSummary): Signal data object with title, type, rank, edge, price, dates, and other fields Returns: Any	`Any`
`run_sample`	Loads a server-local sample OHLCV file, calls the Parallax API, prints trading signals and metadata, and also returns a structured summary object with ok, ticker, date range, CRC, processing time, and signals. This expects a filesystem path visible to the Parallax server process and is mainly for bundled samples or debugging, not workspace-relative files. Params: filepath (str): path to sample OHLCV text file, e.g., ‘data/samples/sample_spy_daily.txt’ Returns: Any	`Any`
`build_request`	Builds an EhSignalRequest payload for daily bars to query ExtremeHurst signals. Params: Inputs: ticker (security symbol), timestamps, opens, highs, lows, closes, volumes (OHLCV arrays) Returns: Any	`Any`

Examples

Run the end-to-end workflow

{
  "request": {
    "ticker": "MSFT",
    "start_date": "2025-01-01",
    "end_date": "2025-12-31",
    "source": "yahoo",
    "refresh": true,
    "last_signal_only": true,
    "history_bars": 252
  }
}

Analyze an existing workspace file

{
  "request": {
    "path": "parallax/daily/MSFT/MSFT_2025-01-01_2025-12-31.txt",
    "last_signal_only": true,
    "history_bars": 252
  }
}

Notes

The composite workflow uses Dataloader for OHLCV retrieval, then runs the signal engine locally.
If protected endpoints are enabled, configure bearer auth in the client or gateway that calls this server.

Client setup

Configure this endpoint in Cursor, Claude Desktop, or a generic MCP client.

Shared tools

Use health, result, artifact, environment, and table helper tools.

Other Servers

Dataloader

Market-data retrieval, TA-Lib indicators, and dataframe exports.

Fama-French Replicate

Official and replicated Fama-French factors plus loadings and alpha estimation.

Statistical Factor Models

Stock-Watson, complete-panel, and dynamic statistical factor extraction.

Jump Models

JumpModel and SparseJumpModel regime fitting, online prediction, and backtesting.

Wavelet Mean Reversion

Wavelet-based mean reversion analysis for financial time series.

EP Ratio Screener

Fundamental stock screening based on earnings yield and balance-sheet quality.

Volatility Scaling Lab

Volatility targeting, EWMA volatility, Monte Carlo bands, and risk diagnostics.

Overview

MCP Setup

Dataloader Server

Fama-French Replicate Server

Statistical Factor Models Server

Jump Models Server

Wavelet Mean Reversion Server