# XBTFX Launches MCP Server and Skills Hub for AI-Connected Forex and CFD Trading

> XBTFX's MCP server and Skills Hub connect compatible AI agents to live forex and crypto CFD accounts across 400+ instruments — no extra subscription fee.

**Published:** 2026-07-01  
**Category:** News  
**Author:** XBTFX Research  
**Canonical:** https://xbtfx.com/blog/xbtfx-launches-mcp-server-and-skills-hub/

---

### **New agent integration layer allows compatible AI assistants, coding agents, and automation frameworks to interact with eligible live XBTFX trading accounts.**

Following the recent launch of our REST and WebSocket Trading API, today we announced the release of an [MCP server and Skills Hub](https://xbtfx.com/page/ai-trading/) designed to connect compatible AI agents and automation frameworks to live forex and CFD trading accounts.

The new agent integration layer allows users to retrieve account information, inspect positions, access market data, and submit structured trading requests through AI tools and development environments they already use.

XBTFX is not launching a proprietary AI model, predefined trading strategy, or black-box trading bot. Traders, strategy builders, and developers decide how the infrastructure is used through natural-language instructions, agent-based workflows, automation frameworks, or custom strategies written in any programming language capable of calling the XBTFX API. Users define the trading logic, instructions, and risk controls; we provide the authenticated account, market-data, and execution infrastructure.

*“Retail trading is moving into a landscape where programming and automation are no longer limited to providers, professional developers and institutions,”*said an XBTFX spokesperson.*“AI now allows traders to turn their own ideas into strategies, monitoring tools, and automated or human-directed workflows. XBTFX is addressing the other side of that equation by making the trading account programmable - giving users structured access to account data, market information, and live trading through the tools and models they choose to build with.”*

## **From Trading API to Agent-Ready Infrastructure**

In June, we launched [Trading API](https://xbtfx.com/page/trading-api/), providing REST and WebSocket access for developers building trading applications, automated strategies, dashboards, reporting tools, and account-management systems.

The MCP server and Skills Hub build on that foundation by presenting the same account, market-data, and trading capabilities in formats that compatible AI agents can understand and use.

The two primary connection paths are:

**MCP-compatible client → XBTFX MCP server → XBTFX Trading API → eligible XBTFX trading account**

**Agent or custom application using XBTFX skills → XBTFX Trading API → eligible XBTFX trading account**

In the first path, the MCP server operates as the structured tool layer between the compatible AI client and the XBTFX Trading API.

In the second, the Skills Hub provides reusable instructions, endpoint definitions, parameters, and examples that help an agent or developer call the Trading API directly. The Skills Hub does not sit within the request path.

In either case, the AI agent interprets the user’s instruction or workflow and selects the appropriate structured function or endpoint. The XBTFX infrastructure then authenticates and processes the corresponding account, market-data, or trading request.

## **MCP Support for Compatible AI Clients**

The XBTFX MCP server exposes Trading API functions as structured tools using the Model Context Protocol.

MCP allows a compatible AI client to discover available tools, understand their required parameters, and invoke them through a standardized interface.

To connect, a user generates an account-bound API key through the [XBTFX Console](https://console.xbtfx.com/), installs or invokes the XBTFX MCP server, and adds the required configuration to a supported client.

Once configured, the agent can access functions including:

- Checking account balance, equity, margin, and floating profit or loss
- Reviewing open positions, pending orders, and trading history
- Retrieving available instruments and contract specifications
- Accessing current market prices
- Opening buy or sell positions with defined volume and risk parameters
- Adding or modifying stop-loss and take-profit levels
- Closing positions fully or partially
- Reversing positions
- Closing positions by instrument
- Monitoring account and market conditions through structured workflows

For example, a user may ask an agent to summarize current account exposure, retrieve the latest price for an instrument, calculate a position size according to defined rules, or submit a specific order with a stop-loss and take-profit.

The natural-language instruction itself is not sent directly to the trading account. The AI client interprets the request and converts it into a structured tool call containing the required account or trading parameters.

The XBTFX MCP server has been tested with environments including:

- Claude Desktop
- Claude Code
- Cursor
- Windsurf
- OpenAI Codex

Other applications may also be compatible where they support the required Model Context Protocol configuration and transport.

## **Skills Hub for OpenClaw, LangChain, CrewAI, and Custom Agents**

For agent environments that do not connect through MCP directly, we have also released a Skills Hub containing reusable instructions and integration resources for the Trading API.

The Skills Hub includes structured files covering:

- Trading operations
- Account information
- Market data
- WebSocket market-data streams

Each skill provides endpoint details, parameter definitions, request examples, expected responses, and operational guidance. This helps an agent understand how to use XBTFX trading functions without requiring a developer to describe every endpoint from the beginning.

The skills can be incorporated into environments and frameworks including:

- OpenClaw
- LangChain
- CrewAI
- Claude Code
- Custom AI-agent implementations

They can also be used by agents powered by models such as Grok, provided the model operates within an environment capable of securely authenticating with and calling the XBTFX Trading API.

This allows developers to choose the integration method that best fits their system. MCP provides standardized tool discovery for compatible clients, while the Skills Hub supports frameworks that use structured instructions, custom tools, or independently managed workflows.

## **Built for Forex, Crypto, and 24/7 Market Workflows**

The agent integration layer is currently available for eligible live XBTFX trading accounts and supports workflows across forex, cryptocurrency CFDs, metals, indices, energies, and stocks.

We also support BTC, ETH, and USDT denominated trading accounts alongside major fiat account currencies. This gives traders and developers the ability to build AI-assisted trading, monitoring, and account-management workflows within a crypto-denominated trading environment, while accessing more than 400 instruments across the broader XBTFX product offering. Available instruments and trading conditions depend on the account associated with the API key.

This capability is particularly relevant to cryptocurrency markets, which operate 24 hours a day, seven days a week. When an agent or automated system is configured to run continuously, it can monitor market data, account conditions, margin, and open exposure outside conventional trading hours and respond according to the user’s predefined instructions and risk controls.

The infrastructure supports different levels of automation. A trader may use an AI assistant to retrieve account information, explain current exposure, or prepare a proposed trading instruction for review. A strategy builder or developer may create a more automated system that monitors market and account conditions and acts according to independently defined strategy, validation, and risk rules.

XBTFX does not provide trading recommendations through the MCP server or Skills Hub, and the integration does not determine when a user should buy or sell an instrument.

## **Developed in Collaboration With HuracanAI**

The MCP server, Skills Hub, Trading API, and supporting execution architecture were jointly developed by XBTFX and [HuracanAI](https://huracan.ai/), a financial-technology firm specializing in brokerage infrastructure, liquidity connectivity, trading bridges, APIs, and AI-agent systems.

HuracanAI contributed core engineering across the connectivity and integration layers, working closely with XBTFX to translate its brokerage requirements, account workflows, and trading capabilities into a unified and documented infrastructure stack for custom applications, automated systems, and compatible AI-agent environments.

## **Account-Level Authentication and User Control**

Each XBTFX API key is associated with an individual eligible trading account. The same authentication model applies whether a user connects through the MCP server, uses resources from the Skills Hub, or integrates directly with the REST and WebSocket APIs.

Users can create and manage their API keys through the [XBTFX Console](https://console.xbtfx.com/). A key can be revoked if it is no longer required or if the user believes it may have been exposed.

Users remain responsible for:

- Securing their API credentials
- Reviewing agent instructions and configurations
- Verifying the connected trading account
- Defining strategy and risk limits
- Monitoring account activity
- Controlling access to the computers and systems on which their agents operate

Because the integration operates with live trading accounts, we recommend starting with limited exposure, testing account-information and market-data functions first, and reviewing agent behavior before enabling broader automated trading workflows.

## **Getting Started**

To connect a compatible AI agent or application:

1. [Open an account](https://my.xbtfx.com) or select an eligible live XBTFX trading account.
1. Generate an account-bound API key through the XBTFX Console.
1. Connect the XBTFX MCP server to a supported AI client, install the appropriate Skills Hub resources, or integrate directly with the Trading API.
1. Verify the connected account and available functions.
1. Begin with account-information and market-data requests before submitting live trading instructions.

The MCP server, Skills Hub, and Trading API are available to eligible account holders without a separate API subscription or per-request fee. Normal spreads, commissions, financing charges, and other trading costs applicable to the connected account continue to apply.

Documentation, installation instructions, client configurations, agent skills, and API references are available through the [XBTFX AI Trading Hub](https://xbtfx.com/page/ai-trading/) and Developer Documentation.

[Get Started](https://my.xbtfx.com/)

## **Risk Warning**

*Trading leveraged products, including contracts for difference, involves substantial risk and may not be suitable for all investors. You may lose some or all of your invested capital.*

*AI agents and automated applications may misunderstand instructions, use incorrect parameters, encounter software or connectivity failures, or behave in unexpected ways. Connecting an AI agent to a trading account does not reduce market risk, guarantee execution quality, or improve trading performance.*

*Users are responsible for reviewing their configurations, securing their API credentials, monitoring account activity, and applying appropriate financial and technical risk controls.*

*Nothing in this announcement constitutes financial, investment, or trading advice. Past performance is not indicative of future results.*