Templates / WhatsApp expense tracker with multi-input (text, image & audio)

WhatsApp expense tracker with multi-input (text, image & audio)

Needs changes

by Roshan Ramani ✓ · 1,368 views · created 2025-06-24 · Personal Productivity, AI Chatbot

Needs changes. This template uses 1 thing(s) that n8n 3.0 removes: AI Agent v1 (old agent modes). 1 breaking item(s) found. Static check, last run 2026-10-19.

What to replace before upgrading

Removed in 3.0Replace with
AI Agent v1 (old agent modes)
Message Intent Classifier (v1.9), Parse & Validate Transaction (v1.9), Build Custom Report Query (v1.9), Build Custom Financial Summary (v1.9), simple conversation (v1.9), Build Confirmation Message1 (v1.9), Parse & Validate Transaction (Receipts & Bills) (v1.9), Parse & Validate Transaction(Audio Transcribe) (v1.9)
Latest AI Agent node (Tools Agent). For SQL Agent, use a Postgres or MySQL tool with a recent AI Agent.

Nodes used (15)

CodeStructured Output ParserCalculatorWhatsApp Business CloudAI AgentPostgresOpenAI Chat Model

What it does

Description CashMate – Your AI-Powered WhatsApp Finance Agent Turn WhatsApp into a smart finance assistant that auto-registers you, logs transactions in natural language, extracts data from receipts and voice notes, and delivers instant report summaries—no apps, no charts, just lightning-fast insights in chat. Who is this for? - Personal finance enthusiasts wanting effortless expense tracking - Freelancers & solopreneurs juggling multiple incomes and expenses - Small business owners needing quick bookkeeping on the go - Busy professionals who prefer messaging over apps - Privacy-minded users who host data on their own PostgreSQL What problem does this solve? - Zero onboarding friction: Just send "Hi"—no forms, no sign-ups - No app switching: Track everything right inside WhatsApp - Manual entry eliminated: Natural-language, image, and voice input all auto-parsed - Instant summaries: On-demand report requests—no dashboards to navigate How it works Auto-Registration - New users: "Hi" → Creates your profile in PostgreSQL - Returning users: Bypasses creation if your number already exists Intent Classification routes every message into one of five branches: Reports & Summaries - Triggers on keywords like "today's report," "show May vs June," or "summary." - Returns concise text summary of income, expenses, and net balance. Natural-Language Transactions - Messages like "Give 200 to Mukesh for car repair" → AI extracts date, category, amount, payee → logs it. Receipt OCR - Attach a receipt image → Gemini OCR node reads line-items → AI categorizes and logs. Voice-Driven Logging - Send a voice note → Deepgram node transcribes → AI logs transaction. General Chat & Greetings - "Hi," "Hello," or casual finance questions → Routed to chat branch for greetings or tips. Setup Prerequisites - n8n instance (self-hosted or n8n.cloud, v1.0+) - WhatsApp Business Cloud API credentials - PostgreSQL database (host, port, user, password) - OpenRouter/OpenAI API key for NLP - Gemini API key for OCR - Deepgram API key for voice transcription Quick Start 1. Import CashMate.n8n.json into n8n. 2. Rename nodes to suit your environment. 3. Configure Credentials in n8n's Credentials section—avoid hard-coding keys. 4. Activate workflow and message "Hi" from WhatsApp. 5. Test by sending a sample expense text, image receipt, or voice note. > Note: All detailed setup instructions and deep configuration steps are provided in the sticky notes within the template. How to Customize - Categories & Currencies: Edit parsing logic in the Function nodes. - AI Models: Swap OpenAI/OpenRouter for GPT-4, Claude, or self-hosted alternatives. - Multi-User Support: Extend registration logic to tag team or family accounts. - Additional Features: Add budget alerts, multi-currency support, or bank integrations via Plaid. Example Interactions 1. Text Transaction yaml User: 300 given to James for the coffee CashMate: ✅ Transaction Added: • Date: 2025-06-24 • Category: Coffee & Beverages • Type: Expense • Amount: ₹300.00 • Counterparty: James 2. Receipt Image (OCR) yaml User: sends image of café bill totaling ₹450 CashMate: ✅ Transaction Added: • Date: 2025-06-24 • Category: Coffee & Beverages • Type: Expense • Amount: ₹450.00 • Counterparty: Café Aroma 3. Voice Transaction yaml User: voice note: "Paid 650 rupees for office stationery" CashMate: ✅ Transaction Added: • Date: 2025-06-24 • Category: Office Supplies • Type: Expense • Amount: ₹650.00 • Counterparty: (none)

View the original on n8n.io → Check your own copy

Get the weekly “what broke” email

New templates, newly broken ones, and n8n release changes. One short email a week.

Form not loading? Subscribe here.