Cloudshelf’s stack and architecture are built with efficient team working foremost in mind, as happy engineers results in great deliveries that delight customers.
Cloudshelf’s platform is based on 6 key modules to allow us to scale and brands and retailers to have the best experience
.avif)
Cloudshelf manager provides a no-code interface for non-tech users to create an entire Cloudshelf set up on in-store devices in minutes.
It is a single code base and can be accessed from within eCommerce CMSs such as Shopify, or standalone, with modes for retailers and Cloudshelf support staff.
This is the most visible part of our solution as it drives the displays on the digital devices. A single engine code base is replicated through a global CDN manages the edge computing and caching to ensure instant response even when networks go down.
This acts as the interface between our platform and the outside word through "connectors". We have ready built ones for Shopify and Salesforce, but the API also allows retailers to build their own to any other platform.
Building on Trigger.dev we facilitate offline processing in a manner that maximises parallel processing while respecting external platform rate constraints.
Our models improve the retailer’s experience by reducing the workload through much of the automation of product and image classification, metadata clean-up and filter optimisation. Over time, the extensive product data sets enhance our competitive advantage. We are moving away from custom built models and become consumers of the advantages from LLMs.
Each engine is customised through a set of light-weight configuration files that that gives retailers full flexibility while ensuring enormous scalability for Cloudshelf.
Keeping our clients data secure is critical for Cloudshelf. By design, we store virtually no personal data, and adopt best practices. For this reason, Cloudshelf has been certified as SOC2 complaint every year.
To throw a new acronym into the room our main stack is PNRN-T, a slightly modified PERN, where we use NestJS and Typescript instead of vanilla Express.
We choose this stack to allow our developers to operate all across the platform, design, code and test end-to-end features
Our engineers have selected the following tools to ease development, communication and testing.
