AI-Powered Web Scraper + Outreach Automation (Amoeba Labs)
Project Summary
One-Line Description:
An intelligent automation system that scrapes business websites, generates audit reports, and conducts personalized AI-driven email outreach campaigns.
What It Does:
Query businesses → Scrape & audit websites → Send personalized emails → Handle replies intelligently
Purpose
Problem Solved
- Manual prospecting - Hours spent researching potential clients
- Generic outreach - Copy-paste emails with low response rates
- Reply management - Manually tracking and responding to leads
- Audit creation - Time-consuming website analysis
Our Solution
- Automated Web Scraping - Extract business info from search results
- AI-Generated Audits - Analyze website performance automatically
- Personalized Outreach - GPT-4 writes custom emails for each lead
- Intelligent Reply Handling - AI responds to replies contextually
- Continuous Conversations - Maintains email threads automatically
How It Works
Step-by-Step Workflow
1. User Input / API Trigger
↓
2. Web Scraping via SerpAPI (Google Search)
↓
3. Website Audit + Scoring (SEO, Performance, UX)
↓
4. AI-Generated Outreach Email (GPT-4.1-mini)
↓
5. Email Sending (SMTP via Gmail)
↓
6. IMAP Reply Monitoring
↓
7. AI Reply Analysis & Response
↓
8. Continuous Conversation Loop
Detailed Process
1. User Input / API Trigger
- Frontend sends query + user email to n8n Webhook
- Previously supported: Telegram Voice Input + Whisper transcription
2. Web Scraping via SerpAPI
- Query converted to Google Search
- Extracts: URLs, business names, contact pages
3. Website Audit + Scoring
- Performance metrics (SEO, Core Web Vitals, UX)
- Generates custom audit report (plain text)
4. AI-Generated Outreach Email
- OpenAI creates personalized email using:
- Issues found in audit
- Lead score
- Business category
- Contextual reasoning
- Report attached automatically
5. Email Sending (SMTP)
6. IMAP Email Reply Handling
- Monitors incoming replies
- AI analyzes reply sentiment:
- Interest → Suggest meeting
- Questions → Respond informatively
- Rejection → Professional fallback
7. Continuous Conversation Logic
- Internal marker
[AUTO_REPLY_FROM_AMOEBALABS] prevents duplicates
- Maintains thread continuity
Tools & Technologies
Automation & Orchestration
| Technology | Purpose |
|---|
| n8n | Core workflow engine |
| Docker | Containerization |
| GCP VM Instance | Deployment |
| Webhook API | External trigger |
AI / ML
| Technology | Purpose |
|---|
| OpenAI GPT-4.1-mini | Email writing + reply reasoning |
| Whisper API | Voice transcription module |
| LLM Logic Nodes | Context classification, scoring |
Web Scraping
| Technology | Purpose |
|---|
| SerpAPI | Google Search wrapper |
| Custom Code Nodes | Regex + HTML parsing |
Communication
| Technology | Purpose |
|---|
| SMTP (Gmail) | Sending emails |
| IMAP | Reading replies |
| Discord / Telegram Bot | Team notifications |
Data Processing
- Custom JavaScript Code Nodes
- Merge & Loop Nodes (collecting items)
Reporting
- N8N File Nodes (text file generation)
- Dynamic audit report creation
Potential Extensions
- Meeting Automation - Calendar API integration
- WhatsApp Business API - Reply handling via WhatsApp
- Multi-Channel Input - Replace Telegram with WhatsApp/Discord
- Multi-Language Support - Global outreach campaigns
- CRM Integration - Push leads to HubSpot, Notion, Google Sheets
Key Features
- Intelligent Scraping - Automated business discovery
- AI Audits - Performance & SEO analysis
- Personalized Emails - GPT-4 powered outreach
- Smart Reply Handling - Context-aware responses
- Thread Continuity - Maintains conversation flow
- Voice Input - Whisper transcription support
- Team Notifications - Discord/Telegram alerts