AI Integration

XBTFX provides two paths for integrating AI agents with the trading platform: the Skills Hub for structured agent instructions and the MCP Server for Model Context Protocol-compatible assistants.

Both approaches give agents full access to the Trading API — executing trades, querying account data, streaming market quotes, and managing positions programmatically.

Integration Paths

Skills Hub

The Skills Hub is a collection of structured markdown files (SKILL.md) that teach AI agents how to use specific XBTFX API capabilities. Each skill contains endpoint documentation, parameter schemas, example requests, and behavioral guidelines.

Skills work with any agent framework:

Claude Code — load skills directly with --skill
LangChain — ingest as tool descriptions
CrewAI — use as agent instructions
OpenClaw — attach as skill files
Custom implementations — parse the structured markdown

MCP Server

The MCP Server exposes XBTFX trading capabilities through the Model Context Protocol, a standard for connecting AI assistants to external tools. This is currently under development.

Available Skills

Skill	Description
Trading	Open, close, modify, and reverse positions
Account	Balance, equity, margin, and trade history
Market Data	Symbol specifications and live quotes
WebSocket	Real-time price feeds and position updates

Quick Start

Get an API key from console.xbtfx.com
Clone the Skills Hub repository or point your agent at a SKILL.md file
Set the XBTFX_API_KEY environment variable
Start trading

# Example with Claude Code
claude --skill ./skills/xbtfx/xbtfx-trading/SKILL.md

Authentication

All AI integrations use the same Bearer token authentication as the REST API. API keys are created at console.xbtfx.com and each key binds to a single MT5 login.

Source Repository

The Skills Hub is open source and available at github.com/XBTFX/xbtfx-skills-hub.