⚡ Track expenses via chat with Claude Haiku and Google Sheets

⚡ 108 views · ⚡ Personal Productivity

Description

💰 AI Expense Tracker — Chat to Track Spending Instantly

Track your expenses by chatting naturally. No forms, no apps — just type and it’s saved.

📖 Description

This workflow turns a simple chat interface into a powerful personal expense tracker. Just describe your spending in plain language — the AI understands it, categorizes it, and saves it to Google Sheets automatically.

Example inputs the AI understands:

• spent 500 on lunch
• uber 150
• paid 1200 electricity bill
• lunch in feb 25 cost 500 ← handles past dates too
• netflix 499
• $50 hotel booking ← detects currency

No rigid formats. No dropdowns. Just type naturally.

✨ Key Features

• Natural language input — type expenses exactly how you’d say them
• AI-powered parsing — Claude Haiku extracts amount, category, date, currency automatically
• 9 auto-detected categories — Food, Transport, Shopping, Bills, Entertainment, Health, Business, Education, Other
• Multi-currency support — INR, USD, EUR, GBP
• Past date handling — “lunch in feb 25 cost 500” saves to February 2025, not today
• Running monthly total — each row stores the cumulative month total
• Monthly summary — type SUMMARY or summary february for any month
• Works on empty sheet — no errors on first use
• Invalid input handling — friendly error if no amount detected

💬 Commands

🛠 Setup Requirements

1. Google Sheet

Create a new Google Sheet with these exact headers in Row 1:

2. Credentials needed

3. After importing

1. Connect your Anthropic credential to the Claude Haiku node
2. Connect your Google Sheets credential to all sheet nodes
3. Update the Sheet ID in all Google Sheets nodes to point to your sheet
4. Open the workflow chat and type your first expense

🏗 How It Works

You type: "spent 500 on car wash"
        ↓
Detect Intent  →  classified as: expense
        ↓
Read All Expenses  →  loads sheet (works even if empty)
        ↓
Prepare Data  →  calculates existing month total
        ↓
AI Parse Expense (Claude Haiku)
  →  amount: 500
  →  category: Transport
  →  description: Car wash
  →  date: today
  →  currency: INR
        ↓
Parse & Total
  →  derives Month from parsed date
  →  computes new running total
        ↓
Is Valid? (amount > 0 and is_expense = true)
  ✅ YES → Save to Sheet → Reply with confirmation
  ❌ NO  → Ask user to include an amount

Summary flow:

You type: "summary february"
        ↓
Detect Intent  →  classified as: summary
        ↓
Read for Summary  →  loads all rows
        ↓
Build Summary
  →  detects "february" in message
  →  filters rows by February (current year)
  →  calculates total, breakdown by category, daily avg
        ↓
Returns formatted report

📊 Sample Summary Output

📊 March 2026 Report

💳 Total: ₹8,450
📝 Entries: 12
📈 Daily avg: ₹470
🔝 Top: 🍕 Food & Dining

Breakdown:
🍕 Food & Dining: ₹3,200 (38%)
🚗 Transport: ₹1,800 (21%)
💡 Bills & Utilities: ₹1,200 (14%)
🛍️ Shopping: ₹1,050 (12%)
🎬 Entertainment: ₹800 (9%)
🏥 Health: ₹400 (5%)

📂 Auto-Detected Categories

⚙️ Workflow Nodes

🔧 Customisation Ideas

• Add a budget alert — warn when monthly total exceeds a set limit
• Telegram integration — replace chat trigger with Telegram bot
• WhatsApp integration — use Twilio WhatsApp as the input channel
• Weekly digest — add a Schedule Trigger for automatic weekly reports
• Multi-user — store user ID with each row to support team expense tracking
• Export to PDF — generate monthly expense report as a PDF

⚠️ Important Notes

• The Read All Expenses node has Always Output Data enabled — this is required so the flow works on an empty sheet
• Month is derived from the parsed date, not today’s date — so past-dated entries file correctly
• The Total column stores the running month total at the time of each entry — it does not update retroactively if you delete rows

📦 Requirements Summary

• n8n (cloud or self-hosted)
• Anthropic API key (Claude Haiku — very low token usage)
• Google account with Sheets access

Built with n8n · Claude Haiku · Google Sheets By Nirav Gajera

🔗 Nodes Used

Google Sheets, Basic LLM Chain, Anthropic Chat Model, Chat Trigger

📥 Import

Download workflow.json and import into n8n: Workflow menu → Import from File

📖 Importing guide · 🔑 Credential setup