BCRP-MCP

BCRP-MCP is a Model Context Protocol server that provides seamless access to Peru’s economic and financial time series data. It exposes search, exploration, and retrieval capabilities for indicators from the BCRP while conforming to MCP interfaces so you can integrate data into AI agents and analysis workflows.

How to use

Connect to the MCP server from your MCP client (for example, Claude Desktop or an IDE with MCP support). You can use either a remote MCP endpoint or a local server you run on your machine. The remote endpoint makes it easy to get started without local setup, while the local server lets you run the service within your own environment for development and testing.

How to install

Prerequisites: Install Node.js (which includes npm) if you plan to use the remote MCP client integration that relies on npx. You also need uv for running a local MCP server, or you can use the remote MCP URL directly.

To run a remote MCP server URL from a client, no local server installation is needed. Use the remote URL provided and connect via the MCP client configuration.

To run a local MCP server, follow these steps exactly to ensure the server starts correctly.

Additional configuration and notes

HTTP remote MCP server configuration (preferred if you want a quick start): Use the following MCP server entry to point clients to the remote endpoint.

{
  "type": "http",
  "name": "bcrp_mcp_remote",
  "url": "https://bcrp-mcp.onrender.com/mcp",
  "args": []
}

Local server configuration and start

If you want to run the server locally, you can start it with uv using the directory where you cloned the project and the main entry script.

{
  "type": "stdio",
  "name": "bcrp_mcp_local",
  "command": "uv",
  "args": [
    "--directory",
    "/path/to/bcrp-mcp",
    "run",
    "main.py"
  ]
}

Tools and endpoints you can use

From the MCP client, you can search for time series groups, list time series within a group, and retrieve data for a specific code. These operations map to the following MCP tools.

• search_time_series_groups: Search for time series groups using keywords.

• search_time_series_by_group: Find all time series within a specific group and return code-name pairs.

• get_time_series_data: Retrieve time series data for a specific code within a date range.

Examples

You can use prompts to drive data discovery and analysis. For instance, you could ask for how the reference interest rate has evolved over the last year and then fetch the relevant series data to analyze the trend.

Security and access notes

If you are using a remote MCP endpoint, ensure you follow your organization’s security practices for API access, including handling tokens or credentials as required by the MCP client and the remote service.

Available tools

search_time_series_groups

Search for time series groups using one or multiple keywords to locate relevant data categories.

search_time_series_by_group

List all time series within a specific group, returning code-name pairs for easy reference.

get_time_series_data

Retrieve time series data for a given code within a defined start and end date.