Author: Shabbir Poonawala

LLMs (Large Language Models) are all the rage, but they have a major flaw: they hallucinate. That means they make things up as they go along. Not ideal when you’re trying to build a chatbot that answers questions strictly from your own documentation. Enter RAG (Retrieval-Augmented Generation), a game-changing technique that fixes this problem. In this post, I’ll break down how to chunk a PDF, create embeddings, and query them in n8n to build a chatbot that answers from your own knowledge base. Get the full tutorial on building a Vector Store with Pinecone and n8n.

Raw leads are useless. Enriched leads are gold. I’ll show you how to take an email address, ping a data provider API (or scrape via Puppeteer), and push the company size and revenue data back into HubSpot using n8n.

This involves using Clearbit and Apollo to fetch the necessary data and n8n to integrate it with HubSpot.

Want to learn more? Check out our Shahi Raj blog for a step-by-step guide.

I got tired of SSH-ing into my server to clean logs. So I wrote a script. Here is the exact Bash code I use to vacuum logs, prune Docker images, and keep my Debian VM clean automatically every morning at 6:25 AM.

It’s a simple yet effective solution that has saved me countless hours of manual maintenance. Want to see the code? Check it out!

Want to automate your server maintenance? Get in touch with our experts today!

Text inputs are easy. But how do you handle a user sending an .ogg voice note or a PDF agenda? I’ll break down how I use n8n to route binary files through OpenAI Whisper (Speech-to-Text) to extract structured meeting data from any format. With ShahiRaj‘s AI Scheduler, I can handle any type of input and automate the meeting scheduling process.

Want to learn more about how I built this multi-modal AI system? Check out my in-depth guide on using n8n to handle voice and file inputs.

#ShahiRaj #n8n #MultiModalAI

Most chatbots are dumb—they just give static answers.

I’ll show you how to give your bot a ‘brain’ by connecting n8n to a MySQL database. When a user asks for an SKU, the bot runs a SELECT query and returns real-time data.

Check out the Shahi Raj blog for more technical insights on chatbot development.

#ShahiRaj #ChatbotDevelopment #n8n

Email is reliable for storage, but it’s terrible for alerts. I’ll show you how to use the IMAP node in n8n to listen for specific ‘High Priority’ emails (like server alerts or client approvals) and instantly route them to a WhatsApp Business API template message.

Here’s a step-by-step guide to building a notification router in n8n:

1. Set up an IMAP node to listen for specific email addresses or keywords.
2. Use a conditional node to filter out low-priority emails.
3. Route the high-priority emails to a WhatsApp Business API template message.
4. Test your setup to ensure everything is working smoothly.

By following these steps, you can create a notification router that sends critical emails to your clients via WhatsApp, ensuring they stay informed and engaged.

Read more about n8n and WhatsApp integrations and how to take your notifications to the next level.

Tags: #n8n #WhatsApp #NotificationRouter #Automation

You shouldn’t send WhatsApp marketing messages at 3 AM on Sunday. I’ll show you how to use the ‘Wait’ node and ‘Time Check’ logic in n8n to queue weekend events and release them in a batch when business opens on Monday.

By implementing a ‘business hours’ logic in your automation workflow, you can ensure that your marketing messages are delivered at the right time. This is crucial for maximizing engagement and conversion rates. In this tutorial, I’ll walk you through the steps to set up a ‘Wait’ node and ‘Time Check’ logic in n8n.

Follow along and learn how to:

• Queue weekend events and store them safely
• Release them in a batch when business opens on Monday

Don’t let your marketing messages go to waste. Implement ‘business hours’ logic in your automation workflow today.

#ShahiRaj #AutomationWorkflows #n8n

Building a dashboard is hard, but building a notification is easy. I’ll show you how I pull data from the Stripe API, Google Analytics, and my Postgres DB, format it into a clean HTML message, and push it to a Telegram Channel via a Cron trigger every Monday at 9 AM.

Want to see the code? Check out my GitHub repository for the n8n workflow.

#ShahiRaj #n8n #Automation

Handling live audio streams is scary. But, I’ll break down the architecture: Exotel (SIP) -> Whisper (Transcribe) -> OpenAI (Decide) -> n8n (Action). How to build a bot that listens and reacts in under 1.5 seconds. Check out our blog for more technical insights and get in touch if you want to build a similar setup.

Building a Voicebot that can handle live audio streams is a complex task, but it’s achievable. If you’re interested in learning more, subscribe to our newsletter and stay updated on the latest tech trends.

#ShahiRaj #n8n #VoiceBot #Technical #Exotel #Whisper #OpenAI

Traditional code breaks when a user says ‘Meet me next Tuesday after lunch.’ I’ll show you how I use Large Language Models (LLMs) to convert fuzzy natural language into strict ISO-8601 timestamps for the Google Calendar API.
With AIMeetGTBot, I can handle ambiguous inputs like ‘next week’ or ‘2 o’clock’. It’s all about contextual understanding and intent recognition. Check out the code and see how you can apply this technique in your own projects.