YNAB FAQs

Question 1

Can LLMs make changes to my YNAB budget using this tool?

Accepted Answer

By default, no. Write tools for creating transactions or updating categories are strictly opt-in, requiring you to set the `YNAB_ALLOW_WRITES=1` environment variable. Even when enabled, these actions are secured with confirmations and amount limits to prevent accidental or malicious changes.

Question 2

What is mcp-ynab and how does it connect with LLMs?

Accepted Answer

mcp-ynab is a Model Context Protocol server that securely exposes your YNAB budgeting data to LLMs (Large Language Models) like Claude. This enables AI clients to inspect your financial plans and transactions, offering advanced financial analysis and insights without direct access to the YNAB API.

Question 3

How does mcp-ynab prioritize data security?

Accepted Answer

Security is paramount. mcp-ynab is read-only by default, preventing unintended mutations. It stores YNAB personal access tokens securely using your OS-native keyring. Optional write capabilities are protected by environment variable gates, elicitation confirmations, amount safety caps, host-locked HTTP requests, and rate limiting.

Question 4

What types of financial insights can LLMs gain from mcp-ynab?

Accepted Answer

LLMs can access comprehensive read-only data including account balances, category spending, individual transactions, monthly summaries, scheduled transactions, and payees. The tool also offers 'task-shaped' functionalities for debt analysis, spending checks, weekly financial reviews, and Sunday ritual dashboards.

Question 5

Which LLM clients are supported by mcp-ynab?

Accepted Answer

mcp-ynab is designed to be compatible with any Model Context Protocol (MCP) client. This includes popular clients like Claude Desktop, Cursor, and Claude Code, allowing for seamless integration into your preferred AI workflow.

YNAB

YNAB

主な機能

ユースケース

主な機能

ユースケース