TrueLayer MCP server for AI assistants

An experimental project connecting TrueLayer APIs and Claude AI

Overview

The Truelayer Model Context Protocol (MCP) server is an experimental project aimed at allowing you to integrate TrueLayer APIs functionalities with an AI assistant.

A popular example of an AI assistant is Claude AI. By integrating the TrueLayer MCP server with this AI assistant, you'd empower Claude AI to perform a range of banking and payment operations on your behalf.

🚧

Developer preview and disclaimer

  • This project is for demonstration purposes only. It is not officially supported or maintained by TrueLayer, does not form a part of the services we contractually provide you with and we are not responsible for the outcome of your use.
  • Use at your own risk.
  • This integration may break with API changes.
  • By using this MCP server with an external LLM you are trusting an external third party with the handling of the data retrieved from the TrueLayer API.

Features

The integration allows an AI assistant to execute some of the TrueLayer actions directly from the assistant chat.

These functions allow you to perform common payment operations like retrieving account information, creating payment links for customers, sending payouts to beneficiaries, and reviewing transaction history.

Given that AI assistants may not always produce deterministic behaviors, and for safe and isolated testing, we recommend testing the MCP server in the TrueLayer Sandbox environment.

Here are specific TrueLayer functions that an AI assistant can call for you once the MCP server is installed:

Account Information

  • get-merchant-account - Retrieve your merchant account details

Payments

  • get-truelayer-payment - Look up details of a specific payment using its ID
  • create-truelayer-payment-link - Generate a new payment link for initiating a payment from a bank account to the TrueLayer merchant account
  • create-truelayer-payout - Initiate a payout from TrueLayer merchant account to a specific external beneficiary (bank account)

Transactions

  • list_transactions - Get transaction history from TrueLayer merchant account with date filtering

Setup Guide for Claude AI integration

Claude AI is a popular AI assistant on which we've run the MCP server installation.

See the TrueLayer MCP Integration for Claude AI repository and follow the installation instructions from the README.md file there to get the TrueLayer MCP server running and integrated with Claude Desktop.

📘

Data privacy

Beware that you are in this way allowing the AI assistant to interact directly with TrueLayer, and sending TrueLayer data to a third party, in this case to Anthropic/ClaudeAI.

By using this your are trusting Anthropic with the processing and storage of data that is retrieved from our APIs. For more information about Anthropic privacy see here: https://privacy.anthropic.com/en/collections/10672565-data-handling-retention


Prerequisites

Ensure you have the following installed and configured:

  • Node.js: Version 14 or later.
  • Package Manager: npm (comes with Node.js) or yarn.
  • TrueLayer Account: An active TrueLayer developer account with API credentials (Client ID, Client Secret, Key ID, Private Key, Merchant Account ID).

Use LLMs for your TrueLayer integration

We want developers integrating TrueLayer to get right resources and being advocated for best practices when they incorporate LLMs into their development workflow.

Is now then possible to access Documentation in plain Markdown.
You can typically access this version by appending .md to the URL of any documentation page. For instance, if you are viewing a guide on our Payments API, you might find its Markdown counterpart at a similar URL ending in .md.
For eg. https://docs.truelayer.com/docs/create-a-payment.md.

This direct Markdown access offers several advantages for AI-driven development, including better guidance for AI Agents.

We also published a TrueLayer docs /llms.txt file, located at https://docs.truelayer.com/llms.txt.

This file acts as a guide for AI tools and agents, instructing them on how to best retrieve and interpret the plain text versions of our documentation.

The adoption of /llms.txt reflects an emerging standard aimed at making web content more accessible and useful for Large Language Models, facilitating a smoother and more accurate integration experience.