AI-powered Amazon product recommendations with Gemini 2.5 Flash via Telegram

Go to Workflow
281 views
Built by Khairul Muhtadin Khairul Muhtadin
Created on November 19, 2025

Description

Send a product name, get back smart Amazon picks. The Decodo Amazon Product Recommender delivers AI-powered shopping recommendations through Telegram: price comparisons, ratings, sales data, and categorized picks (budget, premium, best value) in under 40 seconds. No more hours of manual research.

Last updated: September 2026.

Why Use This Workflow?

Time Savings: Product research drops from 45+ minutes to under 30 seconds.

Decision Quality: 20+ products compared automatically with AI-curated recommendations.

Zero Manual Work: Full automation from message input to formatted recommendations.

Ideal For

E-commerce Entrepreneurs:** Research competitor products, pricing, and market trends for inventory decisions.
Smart Shoppers & Deal Hunters:** Instant comparisons with sales volume data and discount tracking before buying.
Product Managers & Researchers:** Amazon marketplace positioning, customer sentiment, and pricing ranges for competitive intelligence.

How It Works

Trigger: User sends product name via Telegram (e.g., "iPhone 15 Pro Max case")
AI Validation: Gemini 2.5 Flash extracts core product keywords and validates input authenticity
Data Collection: Decodo API scrapes Amazon search results, extracting prices, ratings, reviews, sales volume, and product URLs
Processing: JavaScript node cleans data, removes duplicates, calculates value scores, and categorizes products (top picks, budget, premium, best value, most popular)
Intelligence Layer: AI generates personalized recommendations with Telegram-optimized markdown formatting, shortened product names, and clean Amazon URLs
Output & Delivery: Formatted recommendations sent to user with categorized options and direct purchase links
Error Handling: Admin notifications via separate Telegram channel for workflow monitoring

Setup Guide

Prerequisites

| Requirement | Type | Purpose |
|-------------|------|---------|
| n8n instance | Essential | Workflow execution platform |
| Decodo Account | Essential | Amazon product data scraping |
| Telegram Bot Token | Essential | Chat interface for user interactions |
| Google Gemini API | Essential | AI-powered product validation and recommendations (get a key) |
| Telegram Account | Optional | Admin error notifications |

Installation Steps

Import the JSON file to your n8n instance
Configure credentials:
Decodo API: Sign up at decodo.com → Dashboard → Scraping APIs → Web Advanced → Copy BASIC AUTH TOKEN
Telegram Bot: Message @BotFather on Telegram → /newbot → Copy HTTP API token (format: 123456789:ABCdefGHI...)
Google Gemini: Obtain API key from Google AI Studio for Gemini 2.5 Flash model
Update environment-specific values:
Replace YOUR-CHAT-ID in "Notify Admin" node with your Telegram chat ID for error notifications
Verify Telegram webhook IDs are properly configured
Customize settings:
Adjust AI prompt in "Generate Recommendations" node for different output formats
Set character limits (default: 2500) for Telegram message length
Test execution:
Send test message to your Telegram bot: "iPhone 15 Pro"
Verify processing status messages appear
Confirm recommendations arrive with properly formatted links

Customization Options

Basic Adjustments:
Character Limit**: Modify 2500 in AI prompt to adjust response length (Telegram max: 4096)

Advanced Enhancements:
Multi-language Support**: Add language detection and translation nodes for international users
Price Tracking**: Integrate Google Sheets to log historical prices and trigger alerts on drops
Image Support**: Enable Telegram photo messages with product images from scraping results

Troubleshooting

Common Issues:

| Problem | Cause | Solution |
|---------|-------|----------|
| "No product detected" for valid inputs | AI validation too strict or ambiguous query | Add specific product details (model number, brand) in user input |
| Empty recommendations returned | Decodo API rate limit or Amazon blocking | Wait 60 seconds between requests; verify Decodo account status |
| Telegram message formatting broken | Special characters in product names | Ensure Telegram markdown mode is set to "Markdown" (legacy) not "MarkdownV2" |

Use Case Examples

Scenario 1: E-commerce Store Owner
Challenge: Needs to quickly assess competitor pricing and product positioning for new inventory decisions without spending hours browsing Amazon
Solution: Sends "wireless earbuds" to bot, receives categorized analysis of 20+ products with price ranges ($15-$250), top sellers, and discount opportunities
Result: Identifies $35-$50 price gap in market, sources comparable product, achieves 40% profit margin

Scenario 2: Smart Shopping Enthusiast
Challenge: Wants to buy a laptop backpack but overwhelmed by 200+ Amazon options with varying prices and unclear value propositions
Solution: Messages "laptop backpack" to bot, gets AI recommendations sorted by budget ($30), premium ($50+), best value (highest discount + good ratings), and most popular (by sales volume)
Result: Purchases "Best Value" recommendation with 35% discount, saves $18 and 45 minutes of research time

Quick Answers
How are products grouped?
Budget, premium, best value, and most popular, each with direct purchase links.
What if something fails?
Admin gets notified on a separate Telegram channel for monitoring.
How long are the messages?
2500 characters by default (Telegram max is 4096). Adjustable in the AI prompt.

Made by: khmuhtadin
Category: AI | Productivity | E-commerce | Tags: amazon, telegram, ai, product-research, shopping, automation, gemini

Need custom workflows? Contact me

Connect with the creator:
Portfolio • Workflows • LinkedIn • Medium • Threads

Additional info
Built with n8n. Need an assessment on your business? Feel free to reach out at https://khmuhtadin.com/consultation/

Nodes Used (5)

Basic LLM Chain
@n8n/n8n-nodes-langchain.chainLlm
Code
n8n-nodes-base.code
Google Gemini Chat Model
@n8n/n8n-nodes-langchain.lmChatGoogleGemini
Structured Output Parser
@n8n/n8n-nodes-langchain.outputParserStructured
Telegram
n8n-nodes-base.telegram