AI导哈

个人金融数据分析服务器

一个处理Fineco银行个人金融数据并将其导入本地SQLite数据库的MCP服务器，支持LLM驱动的现金流和投资组合财富分析。

  • by morettimarco
  • Python
  • 本地部署
  • 1 星标
  • 更新于
在 GitHub 查看 morettimarco/fineco-mcp

使用说明

Fineco Bank MCP Server

A Model Context Protocol (MCP) server that ingests personal finance exports from Fineco Bank into a local SQLite database and exposes them as tools for LLM-powered analysis of cash flow and portfolio wealth.

Features

  • Parses Fineco Bank export files (account movements, portfolio snapshot, portfolio orders)
  • Stores data in a local SQLite database with automatic deduplication
  • Supports incremental ingestion — re-run with partial exports and only new records are added
  • Exposes 5 MCP tools consumable by any MCP-compatible client (Claude Desktop, etc.)

Requirements

  • Python 3.10+
  • uv (recommended) or pip

Installation

1. Clone the repository

git clone https://github.com/morettimarco/fineco-mcp.git
cd fineco-mcp

2. Install dependencies

With uv (recommended):

uv pip install -e .

Or with pip:

pip install -e .

This installs mcp[cli], openpyxl, and xlrd.

3. Configure Claude Desktop

Add the server to your Claude Desktop configuration file:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Windows: %APPDATA%\Claude\claude_desktop_config.json
{
  "mcpServers": {
    "fineco": {
      "command": "python3",
      "args": ["/absolute/path/to/fineco-mcp/server.py"]
    }
  }
}

Replace /absolute/path/to/fineco-mcp/ with the actual path where you cloned the repo.

If you used uv, you can use the uv run form instead:

{
  "mcpServers": {
    "fineco": {
      "command": "uv",
      "args": ["run", "python", "/absolute/path/to/fineco-mcp/server.py"]
    }
  }
}

Restart Claude Desktop after editing the config.

Usage

Exporting files from Fineco Bank

Download your exports from the Fineco web platform:

Export type Typical filename Where to find it
Account movements movements_YYYYMMDD.xlsx Conto corrente → Movimenti → Esporta
Portfolio snapshot portafoglio-export.xls Rubrica Titoli → Esporta
Portfolio orders Lista Titoli.xlsx Dossier Titoli → Lista Titoli → Esporta

Note: Despite the .xls/.xlsx extensions, these files are plain comma-separated text (CSV). The server handles this automatically.

Place the files in a folder, then ask Claude:

"Ingest my Fineco exports from ~/Downloads"

Available MCP Tools

Tool Description
ingest_files(directory_path) Scan a folder for Fineco exports and load new records into the database
get_current_liquidity() Return the latest account balance
search_account_transactions(start_date, end_date, keyword, category, limit) Filter and search account movements
get_portfolio_positions() Return current holdings with market value and P&L
search_portfolio_orders(isin, operation_type, sign, limit) Search buy/sell/dividend history

Example prompts

  • "What is my current account balance?"
  • "Show me all expenses in the 'Alimentari' category last month"
  • "How much have I spent on subscriptions in 2025?"
  • "What is my current portfolio allocation by asset class?"
  • "Show me all dividends received for ISIN IE00B4L5Y983"

Incremental ingestion

You do not need to export your full history every time. The server uses SHA-256 hashing on each record's key fields to detect duplicates. Simply export the most recent period and run ingest_files again — only genuinely new records will be added.

Database

The SQLite database (fineco_data.db) is created automatically in the project folder on first run. It contains four tables:

  • account_transactions — all account movements with signed amounts
  • portfolio_orders — full history of buys, sells, and dividends
  • portfolio_snapshot — current portfolio positions (replaced on each snapshot import)
  • balances — timestamped account balance history

Privacy

The database contains personal financial data. It is stored locally only and is excluded from version control via .gitignore. Never commit fineco_data.db or your Fineco export files to a public repository.

相关 MCP