Quickstart

This sections showcases a few ways to get started using our remote MCP server.

---

The Katoshi MCP server can be accessed via this SSE Endpoint: https://server.katoshi.ai/x2/mcp/A17mXBLw/mcp/sse

Authentication is required via Bearer Auth using your Katoshi API Key

The Katoshi MCP server is currently in Beta - Please report any issues or feedback in our Telegram Group

---

n8n

Here's how you can start using the Katoshi MCP with n8n

1) Create a simple new workflow

• Add a simple chat message trigger for testing

  • Can be replaced by webhook triggers (Telegram, Tradingview, etc) or Scheduled triggers to run at fixed intervals.

• Add an AI Agent node

• Add and connect an OpenAI Model node

• Add and connect a Simple Memory node

• Add and connect an MCP Client node

2) Configure the MCP client

• Add the Katoshi MCP SSE endpoint

• Select Bearer Auth

• Add your credential

• Add your Katoshi API Key in the Bearer Token input

4) Configure OpenAI

• Connect your OpenAI API Key in credentials (create one here)
• Select the model of your choice (view all models)

4) Test your Agent

• Click on the bottom Open Chat orange button

• Send a message to your AI agent (e.g. Open a BTC long with $11, 1% sl, 3% tp)

---

General MCP Integrations

Some other MCP clients will ask you for a direct MCP server config file, use the following:

{
  "command": "npx",
  "args": [
    "mcp-remote",
    "https://server.katoshi.ai/x2/mcp/A17mXBLw/mcp/sse",
    "--header",
    "Authorization: Bearer ${API_KEY}"
  ],
  "env": {
    "API_KEY": "********-****-****-************"
  }
}

• Replace the API_KEY env value with your Katoshi API Key

More MCP tutorials coming soon