🤖 Telegram SmartAgent Bot

SmartAgent Bot is an intelligent, AI-driven Telegram bot powered by n8n and Gemini 2.5. It can process both voice and text messages, route tasks to smart agents (like Email, Calendar, Stocks, and more), generate AI-based responses or images, and log everything in Google Sheets.

---

🧠 What It Does

SmartAgent Bot acts as a unified interface for task automation, using AI orchestration and tool-using agents. Users can interact via Telegram using voice or text, and the bot will understand their intent and respond appropriately — whether it's:

• Answering questions via LLM
• Managing emails
• Scheduling calendar events
• Looking up stock prices
• Generating creative prompts or images
• Saving outputs to Google Sheets / Drive

---

🔄 Workflow

---

🔗 Try the Bot Live

You can try SmartAgent Bot directly on Telegram:

👉 @tanishrajput26_bot

---

🏗️ Architecture Overview

1. Message Input

• User sends either a text or voice message via Telegram.
• If it's a voice message, it's transcribed to text using Gemini.

2. Message Type Check (Switch Node)

• The system checks if the user is requesting an image or not.

---

🖼️ If Image Generation is Requested:

• Generate a prompt using Gemini 2.5 Flash.
• Generate a filename using Gemini 2.5 Flash.
• Generate the image using Flux.1-Schnell model.
• Download the image.
• Send the image to the user via Telegram.
• Save the image to Google Drive.
• Log the metadata in Google Sheets.

---

🤖 If Image Generation is Not Requested:

• Query is passed to the AI Orchestrator (Gemini 2.5 Flash).

• Orchestrator intelligently chooses the correct agent:

  • 📧 Email Agent
  • 📅 Calendar Agent
  • 📈 Stock Agent
  • 📊 Growth Agent
  • ✍️ Prompt Agent
  • Or falls back to LLM if no agent is relevant.

• Once the task is processed:

  • Extract clear information for the user.
  • Append the data to Google Sheets.
  • Send a reply via Telegram.
  • If sending fails, send an email notification to you.

---

🧩 Agents & Their Capabilities

Each agent is powered by its own Gemini 2.5 Flash model and has specific tools:

📧 Email Agent

Handles tasks related to Gmail.

• Get Emails
• Label Emails
• Mark as Unread
• Create Draft
• Send Email
• Get Labels
• Reply to Emails

📅 Calendar Agent

Manages events in Google Calendar.

• Create Event
• Create Event with Attendee
• Update Event
• Delete Event
• Get Events

📈 Stock Agent

Fetches live stock data using SerpAPI.

• Web Search Tool (real-time stock prices, market news)

📊 Growth Agent

Finds growth strategies and trends using SerpAPI.

• Web Search Tool (business ideas, market research)

✍️ Prompt Agent

Creative writing, brainstorming, or prompt generation.

• Just LLM access (Gemini 2.5 Flash)

---

🖼️ Image Generation Workflow

When the user asks to generate an image:

1. Prompt Generation → Gemini 2.5 Flash generates a creative prompt.
2. File Name Creation → Another Gemini 2.5 Flash instance generates a Google Drive-safe file name.
3. Image Generation → Uses Flux.1-Schnell model to generate the image.
4. Image Handling:
   • Extract image URL
   • Download image
   • Send it to user in Telegram
   • Save it in Google Drive
   • Log metadata to Google Sheets

---

✅ Features

• 🎙️ Voice-to-text via Gemini
• 🧠 AI-powered Orchestration
• 🧩 Modular Agent Design
• 📤 Auto Logging to Google Sheets
• 🖼️ Image Generation Support
• 📧 Email + 📅 Calendar Automation
• 🌐 Live Web Search via SerpAPI
• ⚠️ Fallback Notifications via Email

---

⚙️ Setup Guide

1. 📦 Requirements

• n8n
• Telegram Bot Token
• Gemini API Key
• Google APIs (Sheets, Drive, Gmail, Calendar)
• SerpAPI Key

2. 📁 Clone the Repo

git clone https://github.com/tanishra/Telegram-Smart-Agent-Bot.git
cd Telegram-Smart-Agent-Bot

3. 🛠️ Environment Setup

Create a .env file based on env.example:

• TELEGRAM_BOT_TOKEN= your_bot_token
• GEMINI_API_KEY= your_gemini_key
• GOOGLE_SHEETS_ID= your_sheet_id
• GOOGLE_DRIVE_FOLDER_ID= folder_id
• SERPAPI_KEY= your_serpapi_key
• YOUR_EMAIL= your_email_for_fallbacks

4. 🔄 Import Workflow into n8n

• Open your n8n instance
• Go to Workflows → Import
• Upload workflows/Telegram_Smart_Agent_Bot.json

5. Start Execution

---

🧪 Testing the Bot

1. Start the Telegram bot and send a message (text or voice).

2. Try these example prompts:

   • 📈 “What is the stock price of Apple?”
   • 📅 “Schedule a meeting tomorrow at 5 PM”
   • 📧 “Reply to this email with thanks”
   • 🖼️ “Generate an image of a robot meditating on a cloud”

3. Observe the outputs:

   • ✅ Response in Telegram chat
   • 📊 Log entry in Google Sheets
   • 📁 Image saved to Google Drive (if applicable)
   • ✉️ Email sent to you in case of failure

---

🧯 Troubleshooting

Issue	Possible Cause	Solution
No response in Telegram	Bot not connected or token invalid	Double-check the TELEGRAM_BOT_TOKEN and ensure the bot is started
Voice not transcribed	Gemini API limit or voice file not accessible	Check Gemini API usage and file permissions
AI Orchestrator returns error	Gemini model failure or bad input	Add validation and fallback handling in the workflow
Email actions fail	Google API scopes or token expired	Re-authenticate Gmail node with correct OAuth scopes
Calendar events not created	Date format or missing fields	Ensure date/time is parsed and formatted correctly
Stock or growth agent returns empty	SerpAPI limit reached or query too vague	Check SerpAPI usage dashboard and refine queries
Image not generated	Prompt too generic or file name issues	Improve prompt clarity and ensure valid file name
File not saved to Drive	Incorrect folder ID or Google Drive API error	Double-check GOOGLE_DRIVE_FOLDER_ID and folder permissions
Sheet row not appended	Google Sheet ID missing or incorrect columns	Ensure the sheet ID is correct and columns match expected format
Telegram message fails to send	Chat ID missing or too large file	Check Telegram node config and reduce message/image size
Fallback email not sent	Email not authenticated or SMTP error	Verify Gmail connection and check fallback node
n8n crashes on load	Too many simultaneous tasks or memory leak	Enable queuing, optimize nodes, increase server memory

---

🤝 Contribution

Contributions, ideas, and improvements are welcome!

🧾 Ways to Contribute

• 🐛 Report bugs via Issues
• ✨ Suggest new agents, tools, or features
• 🛠️ Submit a pull request to improve code or docs
• 📄 Improve the documentation or add diagrams