Ready to create your own intelligent Telegram bot that can remember conversations and perform complex tasks? This comprehensive guide will walk you through building a fully functional AI-powered Telegram bot using three powerful free tools: n8n for workflow automation, Google's Gemini AI for intelligence, and Supabase for persistent memory storage.

What Makes This Bot Special?

Unlike basic chatbots, this AI agent features:

Persistent Memory: Remembers past conversations using Supabase database
AI-Powered Responses: Leverages Google's Gemini model for intelligent interactions1
Tool Integration: Can perform calculations and other specialized tasks
Zero Complex Coding: Built entirely through visual workflow automation
Completely Free: Uses only free tiers of all services

Prerequisites

Before we begin, you'll need:

Telegram account for creating your bot
Google Cloud account for Gemini AI API access
Supabase account for database storage
n8n account (cloud or self-hosted)

Step 1: Setting Up Your Telegram Bot

Create Your Bot with BotFather

The first step is registering your bot with Telegram's official BotFather:

Open Telegram and search for @BotFather
Start a chat and type /newbot
Provide a display name for your bot (e.g., "Smart Assistant")
Choose a unique username ending with "bot" (e.g., "smartassistant_bot")
Copy the access token provided by BotFather1

This token is crucial - it's what allows your n8n workflow to communicate with Telegram.

Step 2: Configure n8n Workflow

Create the Telegram Trigger

In your n8n workspace:

Create a new workflow
Add a Telegram Trigger node
Select "On message" so the workflow activates when users send messages1
Create new credentials using your BotFather token
Test the connection to ensure it's working

Add Google Gemini AI Integration

Next, we'll integrate Google's powerful Gemini AI model:

Add an AI Agent node to your workflow
Select Google Gemini as your chat model1
Create new credentials with your Google AI API key
Configure the connection and test successful authentication1

Step 3: Database Memory with Supabase

Set Up Your Supabase Database

To give your bot long-term memory:

Visit Supabase and create a new project
Generate a secure database password
Choose your preferred region1
Wait for project provisioning to complete

Configure Database Connection

Once your Supabase project is ready:

Navigate to Database Settings → Connect
Copy the transaction pooler connection details1
In n8n, create new Postgres credentials
Paste the host, database name, user string, and password1
Test the connection to confirm it's working

Step 4: Enhance with AI Agent Tools

Add Calculator Tool

To make your bot more capable, add specialized tools:

// The calculator tool allows the AI to perform mathematical calculations

// This is configured directly in the AI Agent node interface

Tool Type: Calculator

Description: "Perform mathematical calculations when requested"

This enhancement allows your bot to handle complex mathematical queries beyond simple conversation1.

Step 5: Configure Message Flow

Set Up Response Handling

The final step connects everything together:

Add another Telegram node for sending responses
Select "Send a text message" action
Configure the chat ID using expression: {{$node["Telegram Trigger"].json["message"]["chat"]["id"]}}
Set message text using: {{$node["AI Agent"].json["output"]}}1

Handle Memory Storage

Critical configuration for conversation memory:

Session ID: Use the chat ID to store conversations per user
Prompt Source: Change from "Connected" to "Define below"
Prompt Expression: {{$node["Telegram Trigger"].json["message"]["text"]}}

Step 6: Testing and Optimization

Activate Your Workflow

Save your complete workflow
Click the toggle to activate it
The workflow is now live and listening

Test Your Bot

Send a message to your Telegram bot and verify:

✅ Receives messages correctly
✅ Processes through AI successfully
✅ Sends coherent responses back to users1
✅ Remembers context from previous messages

Remove n8n Attribution (Optional)

For a professional appearance:

Open the "Send text message" node
Click "Add field" and search for "append attribution"
Toggle off the n8n attribution
Save the workflow

Advanced Features and Use Cases

Real-World Applications

This bot architecture can be adapted for numerous business applications:

Customer Support: Handle common inquiries with context awareness
Personal Assistants: Schedule meetings, set reminders, answer questions
Educational Bots: Provide tutoring with memory of student progress
Business Automation: Process requests and integrate with other systems

Scaling Considerations

As your bot grows, consider:

Database optimization for large conversation histories
Rate limiting to prevent abuse
Advanced AI tools for specific business functions
Multi-language support for global audiences

Benefits of This Architecture

Cost-Effectiveness

Free tier usage across all platforms
No server maintenance with cloud services
Pay-as-you-scale pricing models

Technical Advantages

Visual workflow design eliminates complex coding1
Modular architecture allows easy feature additions
Database persistence enables sophisticated conversation flows
API integrations connect to virtually any service

Troubleshooting Common Issues

Connection Problems

Verify API credentials are correctly entered
Check network connectivity to all services
Confirm service quotas haven't been exceeded

Memory Issues

Validate database schema matches n8n requirements
Test Supabase connection independently
Monitor conversation storage for proper session handling

What's Next?

Your intelligent Telegram bot is now complete and ready for real-world deployment. This foundation provides endless possibilities for expansion:

Add custom tools for specific business functions
Integrate external APIs for enhanced capabilities
Implement advanced AI features like image generation or document processing
Scale across multiple communication platforms

n8n AI Telegram Bot Workflow - JSON Download

Here's the complete n8n workflow in JSON format that you can directly import into your n8n instance to create the AI Telegram bot from the previous guide:

Complete Workflow JSON

{

"nodes": [

{

"parameters": {

"operation": "onMessage",

"credentials": "Telegram Bot",

"event": "message"

},

"id": "Telegram Trigger",

"name": "Telegram Trigger",

"type": "n8n-nodes-base.telegramTrigger",

"typeVersion": 1,

"position": [

250,

300

]

},

{

"parameters": {

"model": "googleGemini",

"credentials": "Google Gemini API",

"promptType": "defineBelow",

"prompt": "{{$node[\"Telegram Trigger\"].json[\"message\"][\"text\"]}}",

"sessionId": "{{$node[\"Telegram Trigger\"].json[\"message\"][\"chat\"][\"id\"]}}",

"tools": [

{

"toolType": "calculator",

"description": "Perform mathematical calculations when requested"

}

]

},

"id": "AI Agent",

"name": "AI Agent",

"type": "n8n-nodes-base.aiAgent",

"typeVersion": 1,

"position": [

600,

300

]

},

{

"parameters": {

"chatId": "={{$node[\"Telegram Trigger\"].json[\"message\"][\"chat\"][\"id\"]}}",

"options": {},

"text": "={{$node[\"AI Agent\"].json[\"output\"]}}",

"operation": "sendMessage"

},

"id": "Send Message",

"name": "Send Message",

"type": "n8n-nodes-base.telegram",

"typeVersion": 1,

"position": [

950,

300

]

}

],

"connections": {

"Telegram Trigger": {

"main": [

[

{

"node": "AI Agent",

"type": "main",

"index": 0

}

]

]

},

"AI Agent": {

"main": [

[

{

"node": "Send Message",

"type": "main",

"index": 0

}

]

]

}

},

"active": true

}

How to Import This Workflow

Step 1: Save the JSON File

Copy the JSON code above
Save it as telegram-ai-bot.json on your computer
Ensure the file extension is .json

Step 2: Import into n8n

Open your n8n instance (cloud or self-hosted)
Click the "+" button to create a new workflow
Go to the three-dot menu (⋯) in the top right
Select "Import from file"
Choose your telegram-ai-bot.json file
Click "Import"

Required Credentials Setup

After importing, you'll need to configure these credentials:

1. Telegram Bot Credentials

Name: "Telegram Bot" (as referenced in the JSON)
Bot Token: Your token from BotFather
Test the connection before proceeding

2. Google Gemini API Credentials

Name: "Google Gemini API" (as referenced in the JSON)
API Key: Your Google AI API key
Model: Ensure Gemini is properly configured

Optional: Add Supabase Memory (Database Configuration)

To add persistent memory as described in the blog post, you would need to:

Add a Postgres node before the AI Agent
Configure Supabase connection:
Host: Your Supabase database host
Database: Your project database name
Username: Your database username
Password: Your database password
Modify the AI Agent to use database context

Testing Your Imported Workflow

Activate the workflow by toggling the switch
Send a message to your Telegram bot
Check the execution log in n8n for any errors
Verify responses are being sent back to Telegram

Workflow Features

This imported workflow includes:

✅ Telegram message listening
✅ AI-powered responses via Google Gemini
✅ Calculator tool integration
✅ Session management using chat ID
✅ Automatic response sending

Troubleshooting Import Issues

If you encounter problems:

Check n8n version compatibility - ensure you're using a recent version
Verify node types exist - some custom AI nodes may need separate installation
Update credentials names if they don't match your existing setup
Test each node individually before activating the full workflow

This JSON file provides a complete, ready-to-use foundation for your AI Telegram bot that anyone can easily import and customize according to their specific needs.