AI-Powered Spending Analysis
Ask Claude about your spending. Get real answers from your actual transactions. No spreadsheets, no exports—just ask.
Powered by get_transactions MCP tool · Read-only access
How it works
Three steps to conversational spending analysis.
Install
Run npx @warmio/mcp to add Warm to your AI client
Connect
Add your API key from Settings → API Keys in Warm
Ask
Ask Claude anything about your spending patterns
The get_transactions Tool
Claude uses this MCP tool to page through your transaction history. Pagination is strict and cursor-based, so the client follows the next cursor until it has the data it needs.
Number of transactions to return per page
Opaque cursor from the previous page of results
Optional knowledge checkpoint for incremental transaction sync
Example conversations
Real questions, real answers from your spending data.
You: "How much did I spend on restaurants last month?"
Claude: You spent $847.32 on restaurants in December across 23 transactions. Your top spots were Chipotle ($156), Thai Basil ($142), and Starbucks ($98). That's about 12% of your total spending.
You: "What are my biggest expense categories?"
Claude: Your top 5 categories this month: Housing ($2,100), Groceries ($623), Transportation ($412), Dining ($387), and Entertainment ($234). Housing takes up 45% of your spending.
You: "Which merchants do I spend the most at?"
Claude: Your top merchants: Amazon ($892), Whole Foods ($456), Costco ($387), Target ($234), and Uber ($198). Amazon leads with 18% of your non-housing spend.
You: "How has my spending changed compared to last month?"
Claude: Your total spending increased 8% from November. Groceries up 15% (+$82), Dining down 12% (-$53), Entertainment up 34% (+$59). The grocery increase is mainly from holiday shopping.
You: "Show me unusual spending this week"
Claude: I spotted 3 transactions above your typical patterns: $287 at Best Buy (usually $0/month in electronics), $156 at a new restaurant, and $89 at an unfamiliar gas station in Nevada.
You: "What percentage of my income goes to food?"
Claude: Based on your deposits, food (groceries + dining) is about 18% of your take-home pay. That breaks down to 11% groceries and 7% dining out.
Why AI spending analysis works
Get the insights you need without the friction.
Ask Anything
No menus or filters to navigate. Just ask "Where is my money going?" in plain English and get a real answer.
Category Breakdowns
See spending by category, merchant, or time period. Claude calculates percentages and totals automatically.
Trend Analysis
Compare spending across months. Spot increases before they become problems. Understand your patterns.
Merchant Insights
Find your top merchants, discover spending patterns, and see which stores get most of your money.
Transaction Details
Drill into specific transactions. Find that one charge you forgot about. Search by amount, date, or description.
Any Time Period
This week, last month, the past year—ask about any timeframe and get accurate totals instantly.
Your data is secure
Built for security from the ground up.
Read-Only
The MCP server can only query data—never modify, delete, or transfer anything.
Scoped Access
Your API key only accesses your accounts. No one else's data.
Instant Revoke
Delete your API key in Settings and access stops immediately.
Common questions
Everything you need to know about AI spending analysis.
How does AI spending analysis work?
Warm's MCP server gives Claude read-only access to your transaction data via the get_transactions tool. When you ask a question, Claude pages through your actual transactions and calculates the answer—no pre-built reports, just real analysis of your data.
What spending questions can I ask?
Anything about your transactions. Category totals, merchant rankings, trend comparisons, unusual charges, percentage breakdowns, time-based analysis—if it involves your spending data, you can ask about it.
Is this different from the spending page in the app?
Yes. The app shows pre-built charts and reports. AI spending analysis lets you ask specific questions in natural language. "How much more did I spend on groceries this month vs my 3-month average?" gets you a direct answer.
How accurate is the analysis?
Claude analyzes your actual transaction data from Warm. The numbers come directly from your connected bank accounts, so accuracy depends on your transaction categorization.
Can Claude see my full transaction history?
Yes. The get_transactions tool is strict and cursor-paginated. Claude can keep following the next cursor across pages when it needs more transaction history for a longer analysis.
Which AI clients support this?
Claude Code, Cursor, Windsurf, Claude Desktop, OpenCode, Codex CLI, Antigravity, and Gemini CLI. Any MCP-compatible client works with the Warm MCP server.
Is my spending data secure?
Yes. The MCP server only has read-only access—it cannot modify transactions, make payments, or change anything. Your API key is scoped to your accounts only, and you can revoke it instantly.
Do I need Pro for this?
Yes. AI spending analysis via MCP is exclusive to Pro ($24.99/mo). Free and Plus tiers do not include API or MCP access.
Ready to analyze your
spending?
Get Pro and start asking Claude about your spending today.
AI features exclusive to Pro. Install with npx @warmio/mcp