Bancoach

Asistente financiero inteligente. El sistema permite a los usuarios consultar balances, metas, etc. mediante una interfaz conectada a un servidor MCP y un LLM.

Comment

Inspiration

Bancoach was born from the need to provide a financial platform that combines artificial intelligence with real-time financial data.
The inspiration came from observing how many people struggle to fully understand their financial situation and make informed financial decisions.
We wanted to create an assistant that not only displays data but also interprets it and guides the user.

What it does

Bancoach allows users to:

  • Check transactions, credits, goals, and personal financial balances.
  • Access macroeconomic indicators such as exchange rates, CETES, and inflation directly from Banxico.
  • Interact with a language model through a web interface to receive personalized financial guidance.

How we built it

The project is composed of three main components:

  1. MCP Server: A Python server connected to Supabase and Banxico, exposing financial tools.
  2. MCP Client API: A FastAPI backend that acts as a bridge between the LLM on OpenRouter and the MCP Server.
  3. Next.js WebApp: A web interface that allows users to interact with the system in a simple and intuitive way.

We used Docker containers for deployment and environment variables for secure configuration.

Challenges we ran into

  • Integrating the communication flow between the LLM, API, and MCP server.
  • Ensuring financial tools were accurate and responded quickly.
  • Avoiding timeout and reconnection issues on free hosting platforms.
  • Managing tokens and credentials securely.

Accomplishments that we're proud of

  • Building a fully functional three-layer architecture in less than 36 hours.
  • Establishing stable communication between a language model and a financial server.
  • Designing an intuitive web interface that translates natural language into real financial actions.

What we learned

  • How to structure a multi-layer system with MCP and LLMs.
  • Best security practices for handling API keys and connections.
  • How to optimize latency between the model, server, and database.
  • The importance of a clear user experience for financial data.

What's next for Bancoach

  • Integrating predictive analysis for expenses and income.
  • Adding personalized investment recommendations.
  • Scaling the infrastructure to more robust cloud services.
  • Developing a native mobile version.
  • Integrating interactive financial education into the platform.

Built With

  • fastmcp
  • next.js
  • openrouter
  • python
  • render
  • supabase
Share this project:

Updates