Track expenses automatically with Telegram Bot using GPT-4o, OCR and voice recognition

Name: Track expenses automatically with Telegram Bot using GPT-4o, OCR and voice recognition
Availability: InStock
Rating: 4.5 (79 reviews)
Author: Dmitrij Zykovic

$20/month : Unlimited workflows

2500 executions/month

Try free

THE #1 IN WEB SCRAPING

Scrape any website without limits

Try free

HOSTINGER 🎉 Early Black Friday Deal
DISCOUNT 20%

Self-hosted n8n

Unlimited workflows - from $4.99/mo

Try free

#1 hub for scraping, AI & automation

6000+ actors - $5 credits/mo

Try free

Personal Expense Tracker Bot 💰

AI-powered Telegram bot for effortless expense tracking. Send receipts, voice messages, or text - the bot automatically extracts and categorizes your expenses.

✨ Key Features

📸 Receipt & Invoice OCR - Send photos of receipts or PDF invoices, AI extracts expense data automatically
🎤 Voice Messages - Speak your expenses naturally, audio is transcribed and processed
💬 Natural Language - Just type "spent 50 on groceries" or any text format
🌍 Multilingual - Processes documents in any language (EN, DE, PT, etc.)
📊 Smart Statistics - Get monthly totals, category breakdowns, multi-month comparisons
🔒 Private & Secure - Single-user authorization, only you can access your data
⚡ Zero Confirmation - Expenses are added instantly, no annoying "confirm?" prompts

🎯 How It Works

Send expense data via Telegram:
- Photo of receipt
- PDF invoice
- Voice message
- Text message
AI processes automatically:
- Extracts amount, date, vendor
- Categorizes expense
- Stores in organized format
Query your expenses:
- "Show my expenses for November"
- "How much did I spend on groceries?"
- "Compare last 3 months"

📋 Expense Categories

Groceries, Transportation, Housing, Utilities, Healthcare, Entertainment, Dining Out, Clothing, Education, Subscriptions, Personal Care, Gifts, Travel, Sports, Other

🔧 Setup Requirements

1. Telegram Bot

Create a Telegram bot via @BotFather and get your API token.

Configure credentials for nodes:

Input, WelcomeMessage,
GetAudioFile, GetAttachedFile, GetAttachedPhoto
ReplyText, NotAuthorizedMessage, DeleteProcessing

2. OpenRouter API

Get API key from OpenRouter for AI processing.

Configure credentials for:

Gpt4o (main processing)
Sonnet45 (expense assistant)

3. Ainoflow API

Get API key from Ainoflow for storage and OCR.

Configure Bearer credentials for:

GetConfig, SaveConfig
ExtractFileText, ExtractImageText
TranscribeRecording
JsonStorageMcp (MCP tool)

🏗️ Workflow Architecture

Section	Description
Message Trigger	Receives all Telegram messages
Bot Privacy	Locks bot to first user, rejects unauthorized access
Chat Message / Audio	Routes text and voice messages to AI
Document / Photo	Extracts text from files via OCR and forwards to AI
Root Agent	Routes messages to Expense Assistant, validates responses
Expense Assistant	Core logic: stores expenses, calculates statistics
Result / Reply	Sends formatted response back to Telegram
Cleanup / Reset	Manual trigger to delete all data (⚠️ use with caution)

💬 Usage Examples

Adding Expenses

📸 [Send receipt photo]
→ Added: 45.50 EUR - Groceries (Lidl)

🎤 "Bought coffee for five euros"  
→ Added: 5.00 EUR - Dining Out (coffee)

💬 "50 uber"
→ Added: 50.00 EUR - Transportation (uber)

Querying Expenses

"Show my expenses"
→ November 2025: 1,250.50 EUR (23 expenses)
   Top: Groceries 450€, Transportation 280€, Dining 220€

"How much on entertainment this month?"
→ Entertainment: 85.00 EUR (3 expenses)

"Compare October and November"  
→ Oct: 980€ | Nov: 1,250€ (+27%)

📦 Data Storage

Expenses are stored in JSON format organized by month (YYYY-MM):

{
  "id": "uuid",
  "amount": 45.50,
  "currency": "EUR",
  "category": "Groceries",
  "description": "Store name",
  "date": "2025-11-10T14:30:00Z",
  "created_at": "2025-11-10T14:35:22Z"
}

⚠️ Important Notes

First user locks the bot - Run /start to claim ownership
Default currency is EUR - AI auto-detects other currencies
Cleanup deletes ALL data - Use manual trigger with caution
No confirmation for adding - Only delete operations ask for confirmation

🛠️ Customization

Change default currency in agent prompts
Add/modify expense categories in ExpenseAssistant
Extend Root Agent with additional assistants
Adjust AI models (swap GPT-4o/Sonnet as needed)

📚 Related Resources

💼 Need Customization?

Want to adapt this template for your specific needs? Custom integrations, additional features, or enterprise deployment?

Contact us at Ainova Systems - We build AI automation solutions for businesses.

Tags: telegram, expense-tracker, ai-agent, ocr, voice-to-text, openrouter, mcp-tools, personal-finance

Dmitrij Zykovic

0 workflows

Nodes

set gmail telegram agent google-gemini

Complexity

advanced

Published 30 Nov 2025

Likes 0

View on n8n.io Download Workflow

✨

Share Your Workflow

Have a great workflow to share? Join the n8n Creator Hub and help the community!

Submit Your Template How to Submit

Related Workflows

Build a Telegram AI assistant with MemMachine, OpenAI, and voice support

# Build a Telegram assistant with MemMachine and voice support An AI assistant that NEVER forgets using MemMachine for persistent cross-session memory, with voice transcription support and productivity tools. **⚠️ Important Deployment Note:** This workflow is designed for **self-hosted n8n** instances. If you're using n8n Cloud, you'll need to deploy MemMachine to a cloud server and update the HTTP Request URLs in nodes 4, 5, and 9. ## What This Template Does This workflow creates an intelligent personal assistant that maintains perfect memory across all conversations, whether you message today or weeks from now. It supports both text and voice messages, automatically transcribes voice using OpenAI Whisper, and provides tools for Gmail, Google Sheets, and Google Calendar. ## Key Features - 🧠 **Perfect Memory** - Remembers every conversation using MemMachine - 🎤 **Voice Transcription** - Supports voice messages via OpenAI Whisper - 📧 **Gmail Integration** - Send and read emails - 📊 **Google Sheets** - Read and write spreadsheet data - 📅 **Google Calendar** - Create and manage events - 🔧 **MCP Tools** - Extensible tool architecture - 💬 **Smart Context** - References past conversations naturally ## Real-World Example **Day 1 - Text Message:** - User: "Send an email to [email protected] about the Q1 report" - AI: *Uses Gmail tool* "Email sent to John about the Q1 report!" **Day 3 - Voice Message:** - 🎤 User: "What did I ask you to do for John?" - AI: "On January 5th, you asked me to email John about the Q1 report, which I sent." **Day 7 - Text Message:** - User: "Follow up with John" - AI: "I'll send a follow-up email to [email protected] about the Q1 report that we discussed on Jan 5th." The AI remembers who John is, what you discussed, and when it happened - all without you having to repeat yourself! ## How It Works ### Message Flow **For Text Messages:** 1. Telegram Trigger receives message 2. Extract user data and message text 3. Store message in MemMachine 4. Search conversation history (last 30 memories) 5. AI processes with full context + tools 6. Store AI response for future reference 7. Send reply to user **For Voice Messages:** 1. Telegram Trigger receives voice message 2. Download voice file 3. OpenAI Whisper transcribes to text 4. Extract transcribed text and user data 5. Store in MemMachine (same as text flow) 6. Process with AI + tools 7. Send reply to user ## Requirements ### Services & Credentials - **MemMachine** - Open-source memory system (self-hosted via Docker) - **Telegram Bot Token** - From @BotFather - **OpenAI API Key** - For AI responses and voice transcription - **Gmail OAuth** - For email integration (optional) - **Google Sheets OAuth** - For spreadsheet access (optional) - **Google Calendar OAuth** - For calendar management (optional) ### Installation ## MemMachine Setup ```bash # Clone and start MemMachine git clone https://github.com/MemMachine/MemMachine cd MemMachine docker-compose up -d # Verify it's running curl http://localhost:8080/health ``` ## Workflow Configuration ### Deployment Options This workflow supports two deployment scenarios: **Option 1: Self-Hosted n8n (Recommended)** - Both n8n and MemMachine run locally - Best for: Personal use, development, testing - Setup: 1. Run MemMachine: `docker-compose up -d` 2. Use `http://host.docker.internal:8080` in HTTP Request nodes (if n8n in Docker) 3. Or use `http://localhost:8080` (if n8n installed directly) **Option 2: n8n Cloud** - n8n hosted by n8n.io, MemMachine on your cloud server - Best for: Production, team collaboration - Setup: 1. Deploy MemMachine to cloud (DigitalOcean, AWS, GCP, etc.) 2. Expose MemMachine via HTTPS with SSL certificate 3. Update HTTP Request URLs in nodes 4, 5, 9 to: `https://your-memmachine-domain.com` 4. Ensure firewall allows n8n Cloud IP addresses ### Configuration Steps 1. **Import this template** into your n8n instance 2. **Update MemMachine URLs** (nodes 4, 5, 9): - **Self-hosted n8n in Docker**: `http://host.docker.internal:8080` - **Self-hosted n8n (direct install)**: `http://localhost:8080` - **n8n Cloud**: `https://your-memmachine-domain.com` 3. **Set Organization IDs** (nodes 4, 5, 9): - Change `your-org-id` to your organization name - Change `your-project-id` to your project name 4. **Add Credentials:** - Telegram Bot Token (node 1) - OpenAI API Key (nodes 4, 7) - Gmail OAuth (Gmail Tool node) - Google Sheets OAuth (Sheets Tool node) - Google Calendar OAuth (Calendar Tool node) ## Use Cases ### Personal Productivity - "Remind me what I worked on last week" - "Schedule a meeting with the team next Tuesday" - "Email Sarah about the proposal" ### Customer Support - AI remembers customer history - References past conversations - Provides contextual support ### Task Management - Track tasks across days/weeks - Remember project details - Follow up on action items ### Email Automation - "Send that email to John" (remembers John's email) - "What emails did I send yesterday?" - "Draft an email to the team" ### Calendar Management - "What's on my calendar tomorrow?" - "Schedule a meeting with Alex at 3pm" - "Cancel my 2pm meeting" ## Customization Guide ### Extend Memory Capacity In **Node 5 (Search Memory)**, adjust: ```json "top_k": 30 // Increase for more context (costs more tokens) ``` ### Modify AI Personality In **Node 7 (AI Agent)**, edit the system prompt to: - Change tone/style - Add domain-specific knowledge - Include company policies - Set behavioral guidelines ### Add More Tools Connect additional n8n tool nodes to the AI Agent: - Notion integration - Slack notifications - Trello/Asana tasks - Database queries - Custom API tools ### Multi-Channel Memory Create similar workflows for: - WhatsApp (same MemMachine instance) - SMS via Twilio (same memory database) - Web chat widget (shared context) All channels can share the same memory by using consistent `customer_email` identifiers! ## Memory Architecture ### Storage Structure Every message is stored with: ```json { "content": "message text", "producer": "[email protected]", "role": "user" or "assistant", "metadata": { "customer_email": "[email protected]", "channel": "telegram", "username": "john_doe", "timestamp": "2026-01-07T12:00:00Z" } } ``` ### Retrieval & Formatting 1. **Search** - Finds relevant memories by customer email 2. **Sort** - Orders chronologically (oldest to newest) 3. **Format** - Presents last 20 messages to AI 4. **Context** - AI uses history to inform responses ## Cost Estimate - **MemMachine**: Free (self-hosted via Docker) - **OpenAI API**: - Text responses: ~$0.001 per message (GPT-4o-mini) - Voice transcription: ~$0.006 per minute (Whisper) - **n8n**: Free (self-hosted) or $20/month (cloud) - **Google APIs**: Free tier available **Monthly estimate for 1,000 messages (mix of text/voice):** - OpenAI: $5-15 - Google APIs: $0 (within free tier) - Total: $5-15/month ## Troubleshooting ### Deployment Issues **n8n Cloud: Can't connect to MemMachine** - Ensure MemMachine is publicly accessible via HTTPS - Check firewall rules allow n8n Cloud IPs - Verify SSL certificate is valid - Test endpoint: `curl https://your-domain.com/health` **Self-Hosted: Can't connect to MemMachine** - Check Docker is running: `docker ps` - Verify URL matches your setup - Test endpoint: `curl http://localhost:8080/health` ### Voice not transcribing - Verify OpenAI API key is valid - Check API key has Whisper access - Test with short voice message first ### AI not remembering - Verify `org_id` and `project_id` match in nodes 4, 5, 9 - Check `customer_email` is consistent - Review node 5 output (are memories retrieved?) ### Tools not working - Verify OAuth credentials are valid - Check required API scopes/permissions - Test tools individually first ## Advanced Features ### Cloud Deployment Guide (For n8n Cloud Users) If you're using n8n Cloud, follow these steps to deploy MemMachine: **1. Choose a Cloud Provider** - DigitalOcean (Droplet: $6/month) - AWS (EC2 t3.micro) - Google Cloud (e2-micro) - Render.com (easiest, free tier available) **2. Deploy MemMachine** For DigitalOcean/AWS/GCP: ```bash # SSH into your server ssh root@your-server-ip # Install Docker curl -fsSL https://get.docker.com -o get-docker.sh sh get-docker.sh # Clone and start MemMachine git clone https://github.com/MemMachine/MemMachine cd MemMachine docker-compose up -d ``` **3. Configure HTTPS (Required for n8n Cloud)** ```bash # Install Caddy for automatic HTTPS apt install caddy # Create Caddyfile cat > /etc/caddy/Caddyfile << 'CADDYEND' your-domain.com { reverse_proxy localhost:8080 } CADDYEND # Start Caddy systemctl start caddy ``` **4. Update Workflow** - In nodes 4, 5, 9, change URL to: `https://your-domain.com` - Remove the `/api/v2/memories` part is already in the path **5. Security Best Practices** - Use environment variables for org_id and project_id - Enable firewall: `ufw allow 80,443/tcp` - Regular backups of MemMachine data - Monitor server resources ### Semantic Memory MemMachine automatically extracts semantic facts from conversations for better recall of important information. ### Chronological Context Memories are sorted by timestamp, not relevance, to maintain natural conversation flow. ### Cross-Session Persistence Unlike session-based chatbots, this assistant remembers across days, weeks, or months. ### Multi-Modal Input Seamlessly handles both text and voice, storing transcriptions alongside text messages. ## Template Information **Author:** David Olusola **Version:** 1.0.0 **Created:** January 2026 ## Support & Resources - **MemMachine Documentation**: https://github.com/MemMachine/MemMachine - **n8n Community**: https://community.n8n.io - **OpenAI Whisper**: https://platform.openai.com/docs/guides/speech-to-text ## Contributing Found a bug or have an improvement? Contribute to the template or share your modifications with the n8n community! --- **Start building your perfect-memory AI assistant today!** 🚀

View

Adjust morning alarms based on weather and train delays using Google Gemini

**Description:** An intelligent alarm system that wakes you up early only when necessary. 🛡️🤖 This workflow monitors your local weather and train status every morning at 5:00 AM. It uses Google Gemini to analyze the situation. If there is heavy rain or a train delay, it sends an emergency alert immediately (and can trigger smart home devices). If everything is normal, it waits until your usual wake-up time to send a calm briefing. ## Key Features: AI Context Awareness: Uses Gemini to intelligently judge "Emergency" situations based on weather descriptions and news headlines. Dynamic Notification: Emergency Mode: Immediate Email alert + Optional SwitchBot trigger (e.g., turn on lights). Normal Mode: Delays notification until your scheduled wake-up time (90 mins later). Targeted Monitoring: Searches specific train lines via Google News RSS. ## How it works: Trigger: Runs daily at 5:00 AM. Fetch: Gets weather from OpenWeatherMap and train news from Google News RSS. Judge: Gemini analyzes the data. Action: Routes the notification based on the status (Emergency/Normal). ## Setup Requirements: Credentials: OpenWeatherMap API, Google Gemini API, Gmail. Config: Open the "1. Configuration" node to set your Location, Train Line, and Email.

View

Track meal nutrition from meal photos with LINE, Google Gemini and Google Sheets

# AI Meal Nutrition Tracker with LINE and Google Sheets ## Who's it for This workflow is designed for health-conscious individuals, fitness enthusiasts, and anyone who wants to track their daily food intake without manual calorie counting. It is best suited for users who want a simple, AI-powered meal logging system that analyzes food photos one at a time and provides instant nutritional feedback via LINE. ## What it does This workflow processes a single meal photo sent via LINE, analyzes it using Google Gemini AI to identify foods and estimate nutritional content, and stores the data in Google Sheets for tracking. The workflow focuses on simplicity and encouragement: it receives a meal image, performs AI-based food recognition, estimates calories and macronutrients, calculates a health score, provides personalized advice, and replies with a detailed nutritional breakdown on LINE. ## How it works 1. A single meal photo is sent to the LINE bot. 2. The workflow is triggered via a LINE webhook. 3. The image file is downloaded and sent to Google Gemini AI for food analysis. 4. The AI identifies foods and estimates nutritional values (calories, protein, carbs, fat, fiber). 5. A health score (1-10) is calculated with personalized improvement tips. 6. The data is appended to Google Sheets for meal history tracking. 7. The image is uploaded to Google Drive for reference. 8. A formatted nutritional report with advice is sent back as a LINE reply. This workflow is intentionally designed to handle **one image per execution**. ## Requirements To use this workflow, you will need: - A LINE Messaging API account - A Google Gemini API key - A Google account with access to Google Sheets and Google Drive - A Google Sheets document with the following column names: - Date - Time - Meal Type - Food Items - Calories - Protein (g) - Carbs (g) - Fat (g) - Fiber (g) - Health Score - Advice - Image URL ## Important limitations - This workflow **does not support multiple images sent in a single message**. - Sending images in quick succession may trigger multiple executions and lead to unexpected results. - Only the first image in an event payload is processed. - Nutritional values are **AI estimates** based on visual analysis and typical serving sizes. - Accuracy depends on image quality, lighting, and food visibility. - This tool should **not replace professional dietary advice**. These limitations are intentional to keep the workflow simple and easy to understand. ## How to set up 1. Create a LINE Messaging API channel and obtain a Channel Access Token. 2. Generate a Google Gemini API key. 3. Update the Config node with your LINE token, Google Sheets ID, Google Drive folder ID, and daily calorie goal. 4. Configure credentials for LINE, Google Gemini, Google Sheets, and Google Drive. 5. Register the n8n webhook URL in your LINE channel settings. 6. Activate the workflow in n8n and test it with a single meal photo. ## How to customize - Modify the AI prompt in the "Analyze Meal with AI" node to support different languages or dietary frameworks (keto, vegan, etc.). - Adjust the daily calorie goal in the Config node to match individual needs. - Add additional nutritional fields such as sodium, sugar, or vitamins. - Replace Google Sheets with a fitness app API or database. - Integrate with other services to send daily/weekly nutrition summaries. --- **Note:** This workflow was tested using real meal photos sent individually via the LINE Messaging API. Nutritional estimates are approximations and may vary from actual values. For accurate dietary tracking, consult a registered dietitian.

View

👨‍💻

Need Custom Automation?

N8N Automation Expert

Specialized in N8N automation, I design custom workflows that connect your tools and automate your processes.