BCB Open Payment Data

You can access open data on Brazilian Central Bank payment methods through an MCP server that lets your AI assistants query transactional data, monthly and quarterly statistics, merchant information, exchange rates, ATM statistics, cardholder details, and PIX/TED/DOC data. This MCP provides a structured, RESTful interface and supports both local and remote deployments so you can integrate payment data into your AI workflows safely and efficiently.

How to use

Connect to the MCP server using a client that supports the MCP protocol. You can run the server locally for testing or deploy it remotely to expose an HTTPS endpoint for ChatGPT or other compatible LLMs. Use natural language prompts to ask for data such as monthly payment method data, quarterly statistics, card transactions, merchant counts, or exchange rates. The available tools expose focused data views and can be called through straightforward REST endpoints or via the standard MCP path in your client.

How to install

Prerequisites: Node.js and npm installed on your machine.

Clone the project, install dependencies, and build the TypeScript sources.

Configuration

The MCP server can be used in two ways: locally with a desktop client or remotely via a hosted REST endpoint.

1. Claude Desktop (local usage): provide a CLI/desktop configuration that points to the local build of the MCP server so Claude can communicate with it directly.

2. ChatGPT (remote usage): deploy the server behind HTTPS and configure ChatGPT to access the REST endpoints exposed by the MCP server. The server provides a standard set of endpoints for discovery and tool invocation.

Deployment notes

Multiple deployment options are supported. You can deploy to Render or Railway for straightforward hosting, or expose a development instance via ngrok during local testing. When deploying to production, ensure the server is reachable over HTTPS and that the base URL is correctly configured in your MCP client.

Examples of use

Ask your AI assistant: ‘What were the PIX data for December 2023?’ or ‘Show quarterly card transactions for Q4 2023.’ You can also query how many establishments were registered in a given quarter or obtain current exchange rates.

Security and access

Use HTTPS for remote deployments and restrict access to endpoints as appropriate for your environment. Do not expose internal test endpoints to public networks without proper authentication and safeguards.

API and endpoints overview

The MCP server exposes REST endpoints compatible with the OpenAI MCP interface for easy integration with compatible LLMs. Typical endpoints include a root information endpoint, a list of tools, tool invocation, health checks, and MCP information. This enables discovery and utilization of all supported data tools.

Available tools

consultar_meios_pagamento_mensal

Query monthly payment method data (PIX, TED, DOC, boletos, etc.) with optional top, skip and OData filter.

consultar_meios_pagamento_trimestral

Query quarterly data for cards and transfers with optional top, skip and OData filter.

consultar_transacoes_cartoes

Query stock and transactions for cards by quarter with optional ordering and filtering.

consultar_estabelecimentos_credenciados

Query the number of accredited establishments for a given quarter.

consultar_taxas_intercambio

Query market exchange rates and related fees for a given quarter.

consultar_taxas_desconto

Query discount rates charged to establishments for a given quarter.

consultar_terminais_atm

Query statistics about ATM terminals for a given quarter.

consultar_portadores_cartao

Query information about cardholders for a given quarter.