Contact person: Business Client

Phone:Show

Email:Show

Location: Ottawa, Canada

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"Summary
Job Description

We already have a working AI chatbot automation built in n8n.

What is already implemented

AI chatbot workflow built in n8n
Chat triggers via messaging channels (Telegram / WhatsApp)
[login to view URL] web scraping tool integrated
AI agent answers user questions based on scraped page content
Basic memory handling inside the AI agent
The chatbot currently scrapes pages on demand (per request)
This part is already done and functional.
What needs to be added / improved

We are looking for an experienced n8n + AI automation specialist to extend this system with a scalable knowledge architecture.

Website crawler & page discovery

Automatically crawl a website starting from a base URL
Identify and collect relevant pages only (exclude media, junk, duplicate, tracking URLs)
Support dynamic structures (blogs, tour pages, category pages, etc.)
Allow configuration for:
depth
include/exclude rules
domain limits

Vector database integration

Store scraped content in a vector database (open to suggestions: Pinecone, Qdrant, Weaviate, Supabase Vector, etc.)
Chunk content properly (semantic chunks, metadata included)
Store:
page URL
page title
content

last updated timestamp
Use vector search for retrieval-augmented answers (RAG) instead of live scraping

Daily (scheduled) re-scraping system

Set up a daily automated job in n8n that:
re-scrapes pages
detects changes
updates vectors only when content has changed
Handle removed pages and updated content correctly
Log scraping status and errors

AI agent integration
Modify the existing AI agent to:
retrieve answers from the vector database
fallback gracefully when information is missing
Ensure no hallucinations
Answers must be strictly based on vector search results

Clean architecture & documentation
Clean, reusable n8n workflows
Clear naming and structure
Basic documentation explaining:
how crawling works
how vectors are updated

how to add a new website later

Required Experience

Strong experience with n8n
Experience building RAG systems
Vector databases (any of the major ones)
AI agents (OpenAI / Claude / Gemini / OpenRouter)
Web scraping best practices
Handling large text content and chunking

Deliverables
Updated n8n workflows
Vector database setup and integration
Daily scraping automation
AI agent updated to use vector search
Short documentation

Pricing & Estimate (IMPORTANT)
Please include in your proposa:
Fixed price estimate for the full implementation
Time estimate (in days)
Brief explanation of:
which vector DB you recommend and why
how you would handle daily updates efficiently
Examples of similar work (if available)" (client-provided description)