AI Systems & FinTech
Finpal - Banking Tools MCP Server & Mobile Application
A conversational banking platform that combines a Model Context Protocol (MCP) server with a cross-platform mobile application, enabling users to access banking services, financial information, and intelligent assistance through a unified AI-powered experience.
*This project was developed during my Internship at Intertech. Obtained permission to showcase the project, but Complete source code is not publicly available due to company policies.
Explore the Project
Problem
Banking customers often need to switch between multiple applications, websites, and customer support channels to access account information, track transfers, check exchange rates, or locate nearby banking services. This fragmented experience creates unnecessary complexity and delays when users need quick financial information.
Solution
The system uses an LLM agent that performs customer-level operations inside a banking application, communicating through an MCP server between the bank's API, custom MCP tools, and the LLM agent. Through natural language interactions and integrated custom banking tools, users can view account details, monitor transfer statuses, access real-time currency rates, discover nearby ATMs, and receive instant support from an AI-powered conversational assistant.
Architecture
Built with a Python-based MCP server powered by FastMCP, exposing banking functionalities through standardized AI-compatible tools. Firebase Authentication provides secure user management, while a React Native and Expo mobile application delivers a seamless cross-platform experience. The system follows a modular service-oriented architecture, allowing banking tools, authentication services, and conversational AI capabilities to operate independently while communicating through the Model Context Protocol.