Use Warm MCP to read your financial data.

Connect your accounts in Warm. Use a supported AI client to review spending, recurring charges, balances, budgets, and goals from your latest synced data.

Set up Warm MCPSee plans and prices

Read data. Approve record changes.

  • Financial context

    get_financial_context returns account information, financial position, recurring items, budgets, goals, and a transaction index where available.

  • Monthly transactions

    get_transactions retrieves an available YYYY-MM month or the fixed ten-day latest period. Ask your AI to check dates and reconcile totals.

  • Approved operations

    Search the operation catalog, describe the exact input, approve it in Warm, then invoke it with the matching approval. Bank money movement is unavailable.

How the connection works

Warm syncs your connected financial accounts. The local @warmio/mcp package runs on your computer, and your AI client starts it when needed. It calls the Warm API using your API key and returns data to the client.

Your AI interprets the results. It can receive that financial data under its own provider's policies. The connector runs locally, but your AI provider may process the data on its servers.

Bank connections cannot move money or write back to your bank. MCP also offers operations on Warm records, with approval for the exact input before execution. Read the permissions guide.

Step 1 of 3
You ask your AI app a question about your money. Your AI app uses the Warm connector to get your data.

Choose your AI client.

Connect a supported AI tool to your Warm account.

Create an API key in Warm under Settings → Export & API. With Node.js >=18, run the installer and reload your client.

Full setup instructions, Troubleshooting, Prompt library

Warm plans differ in how many bank logins you can connect. Compare current prices and bank login limits. Your AI provider may require a separate subscription.

Install the local connector.