LABSCO SUMMARY

Two routes to one endpoint — take the HTTP one unless your client can't.

The HTTP configuration is a URL and nothing else, with the client running the login. The npm bridge exists for clients that only accept a command, and on Windows the README is emphatic about it: install into a fixed directory and point at node with an absolute path rather than relying on npx, which picks the wrong binary.

What it is

A connection to TradeOS trading intelligence. Point your client straight at the Streamable HTTP endpoint and let it handle OAuth, or run the npm package as a local stdio bridge that proxies to the same endpoint.

What you get

Market analysis, ticker search and spread comparison from a plain-language question
Macro context alongside the instrument you asked about
Custom trading agents you create and manage from the conversation
A Claude Code plugin in tradeos-skills/ that adds an analyze command on top of the connection

Requirements

A TradeOS account with MCP API access. The HTTP route needs only the endpoint https://ai.tradeos.xyz/api/agent/mcp/mcp-call and a browser sign-in. The stdio route needs Node.js 18+, npm install @tradeos/tradeos-mcp, and TRADEOS_ACCESS_TOKEN — a JWT the bundled OAuth helper prints for you to paste into the client config.

Setup effort

One command plus a key — npm install @tradeos/tradeos-mcp, then supply credentials

See installation options

USE CASES

01

Getting a read on a ticker without switching apps

Ask in the chat; the analysis and the macro context come back in the same thread.

02

Comparing spreads across instruments

Describe the comparison you want instead of assembling it from separate screens.

03

Running your own trading agent

Create and manage custom agents from the conversation rather than through a separate console.

04

Connecting a client that only speaks stdio

Run the local bridge with an access token; it proxies every call to the same production endpoint.

BEFORE YOU INSTALL

What to know first

Requires

A client that can reach a remote MCP server over HTTP
TRADEOS_ACCESS_TOKEN in your environment

INSTALLATION

Connect your client

Server setup

npm

TERMINAL

npm install @tradeos/tradeos-mcp

Hosted endpoint

ENDPOINT URL

https://ai.tradeos.xyz/api/agent/mcp/mcp-call

Add this hosted endpoint to your MCP client using the credentials listed under Requirements.

MCP Clients

Cursor — HTTP

CURSOR — HTTP · JSON

{
  "mcpServers": {
    "tradeos": {
      "url": "https://ai.tradeos.xyz/api/agent/mcp/mcp-call"
    }
  }
}

~/.cursor/mcp.json (Windows: %USERPROFILE%\.cursor\mcp.json)

Claude Code — Plugin

CLAUDE CODE — PLUGIN · TERMINAL

claude --plugin-dir ./tradeos-skills

Cursor — stdio

CURSOR — STDIO · JSON

{
  "mcpServers": {
    "tradeos-npm": {
      "command": "node",
      "args": [
        "C:/Users/<YOUR_USERNAME>/Downloads/tradeos-mcp-run/node_modules/@tradeos/tradeos-mcp/build/index.js"
      ],
      "env": {
        "TRADEOS_ACCESS_TOKEN": "<YOUR_TRADEOS_ACCESS_TOKEN>"
      }
    }
  }
}

TradeOS AI