Finance

Bare-bones Model Context Protocol server that stores personal finance transactions in ChromaDB and lets Claude run semantic queries against them.

python

0

GitHub Stars

python

Language

7 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "xinrong-meng-my-finance-mcp": {
      "command": "uv",
      "args": [
        "run",
        "my_finance_mcp.py"
      ],
      "env": {
        "MY_FINANCE_MCP_DIR": "/path/to/custom/finance_mcp_dir"
      }
    }
  }
}

You run a bare-bones Model Context Protocol (MCP) server that stores personal finance transactions and lets you query them later using semantic search. Transactions are kept in a vector store for fast retrieval and in a JSON ledger for backup, so your receipts, statements, and portfolios stay searchable and auditable over time.

How to use

You connect an MCP client to the finance MCP server, start the local server, and perform data storage and queries through the available tools. The typical workflow is: upload financial data, store transactions, then ask natural language questions about spending trends, allocations, or other uploaded data.

Available tools you will use through the MCP client include: store_transactions(transactions) to ingest structured data, query_financial_history(query) to run semantic searches, list_transactions(limit=20, offset=0, category=None) to page through the stored ledger, and delete_transactions(indices=None, delete_all=False, confirm=False) to remove entries. You can then navigate results or invoke the management tools to adjust the stored data.

How to install

Prerequisites: you need a Python environment and the uv runner that executes MCP servers.

Step 1: Install the MCP runtime tooling if you don’t have it yet.

Step 2: Copy the finance MCP server to your local workspace and prepare to run it.

Additional sections

Configuration and data location are described here so you can tailor the setup to your environment and storage needs.

Available tools

store_transactions

Ingest structured transactions into the MCP server for persistence and later querying.

query_financial_history

Perform a semantic search across all stored transactions to answer questions about spending, categories, and trends.

list_transactions

Page through the JSON ledger to inspect stored transactions and view their indices.

delete_transactions

Remove specific transactions by index or wipe all data (requires explicit confirmation).