BubbleBrain — AI Chatbot Backend for E-commerce by Serhii LukashBubbleBrain — AI Chatbot Backend for E-commerce by Serhii Lukash

BubbleBrain — AI Chatbot Backend for E-commerce

Serhii Lukash

Serhii Lukash

The Problem

E-commerce stores lose customers every day because no one is available to answer product questions instantly. Support teams can't scale, and generic chatbots give wrong answers that destroy trust.

The Solution

BubbleBrain is a production-ready AI chatbot backend that answers questions from your actual product catalog — no hallucinations, no generic responses. It compares prices in real-time, captures hot leads, and routes them directly to Telegram — all without a support team.

What I Built

A full FastAPI backend with a five-layer AI pipeline:
RAG Engine — OpenAI Embeddings + Pinecone vector search answers questions strictly from the store's own data
Price Comparator — scrapes supplier sites and compares against WooCommerce prices on demand
Lead Pipeline — classifies intent, captures contacts, routes hot leads to Telegram topics automatically
WooCommerce Integration — real-time webhook sync for orders and inventory
Flowise Widget — embeds into any frontend with zero custom UI work

Technical Highlights

88% test coverage · Prometheus metrics · Sentry error tracking · Rate limiting · Structured logging · Docker + Docker Compose · CI-ready · MkDocs documentation

Tech Stack

Python 3.13 · FastAPI · OpenAI GPT-4 · Pinecone · BeautifulSoup · APScheduler · PostgreSQL · SQLite · Docker · Prometheus · Sentry · Poetry · pytest

Results

Zero hallucinations — bot answers only from verified product data
Lead capture fully automated — hot prospects reach sales team in under 60 seconds
Production-ready from day one — monitoring, logging, and error tracking included
Like this project

Posted Jun 20, 2026

AI-powered chatbot backend for e-commerce — RAG pipeline, price comparison, lead generation, and Flowise widget integration in a single production-ready FastAPI service.