Skip to content
MCP server

Stochastly Desktop

Run backtests from a coding agent through the Stochastly desktop app

Runs locallyStatus and history
Install
Pricing
Paid
Platforms
self-hosted, macOS, Windows
Runs
Locally
Version
0.2.2
Listing updated
2 days ago

About Stochastly Desktop

Stochastly Desktop is a paid MCP server for finance and payments. It runs locally on your machine and is started by your AI client.

This package is a small launcher. It connects an MCP client to the server that ships inside the Stochastly desktop app, which must be installed, running and licensed for the Atelier or Portfolio edition. The package itself contains no engine code. Without the app it still lists the 29 tools, each marked as needing the app, and calls return an answer saying which edition is needed.

The site describes a research loop in which an agent searches a node catalogue, composes a graph, runs a backtest and reads the verdict. Tools named include node_catalog, compose_graph, backtest_graph, backtest_status, red_flag_check, adversarial_check, portfolio_combine and export_paper. Long runs can start in the background and be polled by run id, and a session registry counts finished runs.

Tools are classed as reads or writes, with writes stating what they change and passing a second licence check. Driving the app is off until you create a flag file by hand, and an AST self-test checks the read and write classification.

Install

Run in a terminal

claude mcp add desktop -- uvx stochastly-mcp

Add to claude_desktop_config.json

{
  "mcpServers": {
    "desktop": {
      "command": "uvx",
      "args": [
        "stochastly-mcp"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "desktop": {
      "command": "uvx",
      "args": [
        "stochastly-mcp"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "desktop": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "stochastly-mcp"
      ]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "desktop": {
      "command": "uvx",
      "args": [
        "stochastly-mcp"
      ]
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "desktop": {
      "command": "uvx",
      "args": [
        "stochastly-mcp"
      ]
    }
  }
}

Configuration

  • STOCHASTLY_MCP_SCRIPTOptional path to the MCP client script supplied by the installed Stochastly app.

Tools this server provides (8)

Taken from the project documentation; the running server may differ.

node_catalog
Searches the catalogue of graph nodes.
compose_graph
Composes a strategy graph.
backtest_graph
Runs a backtest, optionally in the background.
backtest_status
Reads the status of a background run.
red_flag_check
Checks a result for red flags.
adversarial_check
Runs an adversarial robustness check.
portfolio_combine
Combines strategies into a portfolio.
export_paper
Writes a research paper PDF to disk.

What it does

Graph composition
Builds strategy graphs from a node catalogue.
Background backtests
Starts runs that can be polled by id until a verdict arrives.
Robustness checks
Includes red-flag and adversarial checks on results.
Read and write classes
Each tool is a read or a write, and writes state what they change.
Licence gating
Atelier or Portfolio is required; portfolio_combine needs Portfolio.
Local only
Talks to the installed desktop app rather than a hosted service.

Good for

  • Have an agent compose and run a backtest
  • Check a result for red flags before trusting it
  • Combine strategies into a portfolio
  • Export a research paper from a run

Pricing

Paid

Requires a licensed Stochastly desktop app (Atelier or Portfolio); prices not published

Prices change. Check stochastly.com before you buy.

Stochastly Desktop FAQ

What do I need to run it?

The Stochastly desktop app, installed, running and licensed for Atelier or Portfolio.

How many tools are there?

29.

Can the agent control the app?

Only after you place a .mcp_pilot flag file by hand; no tool can create it.

How do I install the launcher?

pip install stochastly-mcp or run uvx stochastly-mcp.

Is Stochastly Desktop free?

No. Stochastly Desktop is a paid product.

How do I add Stochastly Desktop to Claude Code?

Run this command in your terminal: claude mcp add desktop -- uvx stochastly-mcp

What are the best alternatives to Stochastly Desktop?

The closest alternatives to Stochastly Desktop are Agent-Native Analytics, Google Ads + Meta Ads + SEO/GEO MCP, Solana MCP by Vybe, Tapetide - Indian Stock Market MCP and World Monitor.

Alternatives to Stochastly Desktop

  1. Open-source analytics app where an AI agent builds charts and dashboards from your data

    • Remote
    • Open source
    • 7.7K GitHub stars
    • Analytics and data
  2. Hosted MCP connector for managing Google Ads, Meta Ads, GA4 and Search Console from chat

    • 8 tools
    • Remote
    • Freemium
    • 8.1K GitHub stars
    • Marketing and SEO
  3. Remote MCP server exposing the Vybe Solana API to AI clients and code editors

    • 44 tools
    • Remote
    • 2.8K GitHub stars
    • Finance and payments
  4. Indian NSE and BSE stock market data for AI assistants through 48 MCP tools

    • 17 tools
    • Remote
    • Freemium
    • 150 GitHub stars
    • Finance and payments
  5. World Monitor

    MCP server

    Live geopolitical intelligence dashboard with an API and MCP server for agents

    • Remote
    • Freemium
    • Analytics and data
  6. Budget Tracker

    MCP server

    Local MCP expense ledger that logs receipts and mileage, splits VAT and prepares rebilling

    • 14 tools
    • Remote
    • Freemium
    • 2 GitHub stars
    • Finance and payments