← Back to Projects

Case Study

Nexora CRM

Production-grade multi-tenant SaaS CRM with lead management, sales pipelines, RBAC, and analytics dashboards.

Live DemoArchitectureCase Study
Next.jsTypeScriptFastAPIPythonPostgreSQLJWT

Overview

Nexora CRM is a full-stack, multi-tenant SaaS platform for sales teams to manage leads, move deals through customizable pipelines, and monitor performance via analytics dashboards. A Next.js frontend pairs with a FastAPI backend and PostgreSQL, using JWT authentication and tenant-scoped APIs to keep each organization's data isolated.

Problem Statement

Growing sales teams often outgrow spreadsheets and siloed tools. They need a single CRM that supports multiple organizations on one platform, enforces role-based permissions, and surfaces pipeline health — without one tenant ever accessing another's data.

Solution Architecture

Nexora CRM uses a decoupled architecture: a Next.js (TypeScript) frontend for dashboards, lead views, and pipeline boards communicates with a FastAPI REST API over HTTPS. PostgreSQL stores tenants, users, leads, pipeline stages, and activity records with tenant_id scoping on every query. JWT middleware authenticates requests and encodes role claims used by RBAC guards on both client routes and API endpoints.

┌──────────────────┐     ┌──────────────────┐     ┌─────────────────┐
│   Next.js CRM    │────▶│  FastAPI REST API │────▶│   PostgreSQL    │
│  (TypeScript)    │     │  (JWT + RBAC)     │     │ (Tenant-scoped) │
└──────────────────┘     └──────────────────┘     └─────────────────┘

Technology Stack

  • Next.js
  • TypeScript
  • FastAPI
  • Python
  • PostgreSQL
  • JWT

Key Features

Multi-Tenant Architecture

Each organization operates in an isolated data boundary — users, leads, pipelines, and analytics are scoped by tenant_id at the database and API layers.

JWT Authentication & RBAC

Signed JWTs carry user identity and role claims. Admin, manager, and rep roles control access to settings, team views, and lead assignment workflows.

Lead & Pipeline Management

Capture and qualify leads, assign owners, and move deals through customizable pipeline stages with activity tracking at each step.

Analytics Dashboards

Pipeline conversion, lead source breakdown, and rep performance metrics aggregated server-side for fast dashboard rendering.

RESTful API Design

Tenant-aware endpoints for leads, pipelines, users, and analytics — validated with Pydantic schemas and protected by middleware on every route.

Responsive UI

Mobile-friendly CRM interface with accessible forms, focus states, and layouts that adapt from pipeline boards on desktop to stacked views on smaller screens.

My Role

End-to-end ownership: designed the multi-tenant data model, built the FastAPI backend with JWT auth and RBAC middleware, implemented the Next.js frontend (lead lists, pipeline boards, analytics views), and deployed the live demo on Vercel.

Tech Decisions

  • PostgreSQL over MongoDB — relational schema fits tenant/user/lead/pipeline relationships and supports efficient aggregations for analytics
  • FastAPI + Pydantic — type-safe request validation and auto-generated OpenAPI docs for rapid API iteration
  • JWT with role claims — stateless auth that scales across serverless frontend and API deployments without session stores
  • Tenant_id column scoping — simpler than schema-per-tenant while keeping queries predictable with composite indexes

Multi-Tenancy & Data Isolation

Every table carries a tenant_id foreign key. API middleware extracts the tenant from the JWT and injects it into query filters, preventing cross-tenant reads or writes even if a client sends another organization's ID.

Pipeline Workflows

Pipeline stages are configurable per tenant. Leads transition through stages with timestamped activity logs, enabling conversion-rate analytics and rep-level performance tracking without external BI tools.

Outcome

Shipped a production-ready multi-tenant SaaS CRM end-to-end — live demo with lead management, pipeline tracking, RBAC, and analytics dashboards. The platform demonstrates full-stack SaaS patterns: tenant isolation, secure auth, and data-driven sales workflows in a single deployable application.

Challenges Faced

  • 01Enforcing tenant isolation on every API path without repetitive boilerplate in route handlers
  • 02Balancing RBAC granularity (admin vs manager vs rep) with a maintainable permission model
  • 03Keeping analytics queries performant as lead and activity volume grows per tenant
  • 04Designing pipeline stage transitions that preserve audit history for conversion reporting

Lessons Learned

  • Tenant scoping belongs in middleware — centralizing it prevents accidental data leaks as new endpoints are added
  • Role claims in JWTs simplify frontend route guards but require careful token refresh when roles change
  • Pre-aggregating pipeline metrics on write reduces dashboard load compared to full-table scans at read time
  • Starting with a single shared schema and tenant_id is faster to ship than schema-per-tenant for early-stage SaaS