Convert websites to audio summaries via WhatsApp using GPT and TTS

How It Works

This workflow transforms any webpage into an AI-narrated audio summary delivered via WhatsApp:

1. Receive URL - WhatsApp Trigger captures incoming messages and passes them to URL extraction 2. Extract & validate - Code node extracts URLs using regex and validates format; IF node checks for errors 3. User feedback - Sends either error message ("Please send valid URL") or processing status ("Fetching and analyzing... 10-30 seconds") 4. Fetch webpage - Sub-workflow calls Jina AI Reader (https://r.jina.ai/) to fetch JavaScript-rendered content, bypassing bot blocks 5. Summarize content - GPT-4o-mini processes webpage text in 6000-character chunks, extracts title and generates concise summary 6. Generate audio - OpenAI TTS-1 converts summary text to natural-sounding audio (Opus format for WhatsApp compatibility) 7. Deliver result - WhatsApp node sends audio message back to user with summary

Why Jina AI? Many modern websites (like digibyte.io) require JavaScript to load content. Standard HTTP requests only fetch the initial HTML skeleton ("JavaScript must be enabled"). Jina AI executes JavaScript and returns clean, readable text.

Setup Steps

Time estimate: ~20-25 minutes

1. WhatsApp Business API Setup (10-15 minutes) - Create Meta Developer App - Go to https://developers.facebook.com/, create Business app, add WhatsApp product - Get Phone Number ID - Use Meta's test number or register your own business phone - Generate System User Token - Create at https://business.facebook.com/latest/settings/systemusers (permanent token, no 4-hour expiry) - Configure Webhook - Point to your n8n instance URL, subscribe to "messages" events - Verify business - Meta requires 3-5 verification steps (business, app, phone, system user)

2. Configure n8n Credentials (5 minutes) - OpenAI - Add API key in Credentials → OpenAI (used in 2 places: "Convert Summary to Audio" and "OpenAI Chat Model" in sub-workflow) - WhatsApp OAuth - Add in WhatsApp Trigger node using System User token from step 1 - WhatsApp API - Add in all WhatsApp action nodes (Send Error, Send Processing, Send Audio) using same token