Documentation

MCP Server Overview ​

The Florentine.ai MCP Server lets you integrate natural language querying for your MongoDB data directly into your custom AI Agent or existing AI App.

How it works ​

Questions are forwarded by the AI Agent to the MCP Server, transformed into MongoDB aggregations and the aggregation results are returned to the Agent for further processing.

There are two ways how you can connect an AI agent to the MCP Server: As a local stdio MCP server or as a remote MCP server.

Local MCP server integration ​

You should choose the local MCP server if you are developing your own custom AI agent or if you need to provide additional configuration parameters (like Required Inputs) when using Florentine.

Local MCP servers are usually added to an MCP setup configuration file, which is typically formatted as json:

json
{
  "mcpServers": {
    "florentine": {
      "command": "npx",
      "args": ["-y", "@florentine-ai/mcp", "--mode", "static"],
      "env": {
        "FLORENTINE_TOKEN": "<FLORENTINE_API_KEY>"
      }
    }
  }
}

Remote MCP server integration ​

You should choose the remote MCP server if you are using an existing AI App and you do not need any additional configuration parameters. You authorize the app via OAuth to access Florentine by using this url:

https://mcp.florentine.ai

Available Tools ​

The server provides two tools:

  • florentine_list_collections --> Lists all currently active collections that can be queried. That includes descriptions, keys and type of values.
  • florentine_ask --> Receives a question and returns an aggregation, aggregation result or answer (depending on the Response Types setting).