E-commerce & RetailAI product contentStore scrapingCustom trading algorithm

Bidygo Backoffice

The Bidygo back office: partner stores, scraped products, categories, price offers, AI-generated product content, payouts and commissions in one admin.

Bidygo back office: AI instructions and scraped products
ClientBidygo
LocationArmenia
IndustryE-commerce & Retail
Duration12 months, ongoing

Client

The client is a startup founder focused on B2B and B2C business. Bidygo is a B2B2C platform that helps retailers maximize sales volume by meeting the demand of buyers who always want to pay less and get the product immediately.

Business needs

Build software that lets buyers offer a price for a product and buy it for less, a model that also helps retailers increase their sales volume.

Our solution

gHost Services built the Bidygo back office: one place where the team manages partner stores, products, categories, price offers, orders, returns, payouts and commissions, and where retailers register their shops and connect sales channels.

Bidygo back office store list

How it works

Products come from partner stores and from online and physical non-partner shops. Each product shows its price, store, categories and how many users added it to a wishlist or made a price offer, with price tracking over time.

Scraped products with prices, stores and user actions

Key features

  • Store management with ratings, status and a drag-and-drop order
  • A category tree with subcategories, brands and options
  • Price offers and a custom trading algorithm
  • Orders, returns and a financial summary with payouts and settlements
  • Commissions by category, countries and taxes
  • Sales channel connections: Amazon, eBay, Shopify and more
  • Analytics dashboards and reports
Bidygo category tree

Bidygo AI

Configurable AI instructions generate product content for partner and scraped stores: descriptions, pros and cons, review summaries with a predicted rating, technical characteristics and search keywords.

Bidygo AI instructions

Scraper generator

The team creates and monitors scraping tasks for online stores, with live statuses, retries and auto-refresh, so new shops and their products flow into the catalog automatically.

Scraper generator tasks

Architecture

API gateway

One Node.js 24 and Express 5 gateway is the only HTTP entry point. It checks JWT tokens, validates requests with Joi, rate-limits through Redis, pushes live updates over Socket.io and talks to the services over gRPC.

Services and infrastructure

Node.js gRPC services for products, shops, orders, customers, files, messaging and logging, on PostgreSQL with Sequelize, Redis, MongoDB logs and AWS S3. Docker Compose for every environment; Grafana, Loki and Promtail for monitoring.

AI and scraping

OpenAI and Google Gemini generate product content, with Qdrant as the vector database. Puppeteer clusters scrape partner and non-partner stores, rate-limited with Bottleneck.

Back office

A React 19 app on Vite with Redux Toolkit, React Router 7, Tailwind CSS and Sass, ApexCharts dashboards, drag-and-drop category trees and a rich text editor.

Challenges

  • A workflow for offering the best price and buying the product you love
  • Syncing sales channels and unifying their categories and filters
  • Selling areas, so sellers can set different prices and currencies in different regions
  • An advanced delivery workflow that shows the right shipping price for each product

Results

A smooth, intuitive and fast UI/UX
Features designed for future growth

Technologies

Node.jsExpressMicroservicesSocket.ioJWTSequelizePostgreSQLRedisMongoDBAWSFirebaseOpenAIGoogle GeminiQdrantPuppeteerReactViteReduxReact RouterTailwind CSSSassDockerGrafanaESLint
Next case study

OTC

→