CertsLibrary Multi-tenant SaaS Platform

CertsLibrary Multi-tenant SaaS Platform · Next.js · TypeScript · OpenAI API

CertsLibrary is a live certification preparation platform serving students studying for IT and professional exams across vendors including Microsoft, Amazon, and Salesforce. Beyond static study materials, the platform includes a fully custom web-based exam engine - timed tests, security enforcement, real-time progress tracking, and post-exam analytics - all built to feel like a real exam environment.

The brief had two distinct layers. The first was a standard SaaS problem: multiple client organizations needed isolated data environments, custom branding, and separate user configurations - all running from one codebase. Get that architecture wrong early and you spend months untangling it.

The second layer was harder: a fully functional browser-based exam engine. Not just rendering questions - a timed, secured, proctored test environment with no tab switching, no developer tools, no right-click, and automatic submission on timeout. That behavior had to be reliable enough that users trust it with a real certification attempt.

Both systems had to work together, and both had to be tenant-aware.

Each exam has a dedicated page showing question count, last updated date, customer count, and star rating - all live data. The pricing system runs three tiers per exam: PDF download, web-based test engine, and AI-powered test engine. Each tier has its own CTA, its own demo/trial path, and its own cart integration. That's a non-trivial component to build cleanly without hardcoding per-exam logic.

Before any test starts, users configure their session - question types, pass percentage (adjustable via slider), time limits, and navigation mode (forward-only or free navigation). The security rules are surfaced clearly upfront: no browser navigation, no tab switching, no developer tools, no right-click, no text selection, no minimizing. Auto-submit fires when time expires. This screen sets expectations and handles all the session parameters that the exam engine consumes.

The exam interface is a three-panel layout: a left sidebar for question navigation with live status indicators (answered, unanswered, current), a main panel for the active question and MCQ options, and a right sidebar showing the running timer, progress ratio, pass percentage target, navigation mode, and real-time answer status counts.

Security behavior runs at the JavaScript level - browser navigation disabled, tab-switch detection ends the exam, developer tools blocked, right-click disabled, text selection disabled. These constraints are active from the moment the exam starts and removed cleanly on completion or timeout.

Post-exam, users land on a detailed results page showing their score against the pass threshold, a breakdown of correct/wrong/skipped/attempted counts, and three charts - score comparison against the pass requirement, questions overview, and time breakdown (total available vs. time taken vs. time remaining). Below the charts, a question-by-question review table shows each answer alongside the correct answer, result status, and a view button to revisit the explanation.

This screen has a lot of data to display without overwhelming the user. The layout keeps the headline score visible at a glance, progressively reveals more detail as you scroll, and uses color consistently (red for wrong, green for correct) to make the review scannable.

Underneath all of this is the multi-tenancy layer - middleware that identifies the active organization on each request and loads tenant-specific configuration before anything renders. Each tenant operates with isolated data, its own branding, its own feature flags, and its own user roles. A new tenant inherits the full platform - exam engine, pricing system, analytics, AI features - without touching the core codebase.

The component library was built to be configuration-driven. Every component accepts tenant-level props for color scheme, layout variants, and content shape. Build once, configure per tenant. That approach kept design 100% consistent across implementations and cut component development time by 50%.

Live platform serving real users across multiple certification vendors

3-tier revenue model implemented - PDF, Test Engine, AI Test Engine - with individual cart and checkout flows per tier

Full browser-based exam engine with timer, security enforcement, and auto-submission

Post-exam analytics with score comparison charts, question breakdown, and answer review

175+ paying customers and a 5.0/5.0 rating visible on the live platform

100% design consistency across tenant implementations from one shared component library

50% reduction in component development time through reusable, config-driven architecture

OpenAI API integration powering the AI Test Engine tier