BackendCompleted

Shortify - URL Shortener

Nov 2025

PythonFastAPIMongoDBRedisDocker

Shortify started as a way to practice system design beyond basic CRUD, so most of the interesting work happens on the redirect path. Hashing, storage, and caching are all tuned to keep lookups fast as the URL table grows.

How It Works

Incoming URLs get a short ID from a SHA-256 hash of the long URL plus user ID, encoded in Base62 for a compact, URL-safe string. If that ID already exists, a UUID salt gets appended and the URL is re-hashed, which keeps the scheme deterministic without needing a central counter.
MongoDB holds the long-term URL mappings. It was picked for its flexible document model and the ability to scale horizontally as the dataset grows.
A three-tier Redis cache sits in front of MongoDB with a 7-day TTL that resets on every access, so hot redirects resolve in well under a millisecond without ever touching the database.

Request → Redis (Cache) → MongoDB (Source of Truth)
          ↓ miss              ↓ found
          Query DB --------→ Cache result

Database calls run through Motor, MongoDB's async driver, so the FastAPI service can handle many concurrent redirects without blocking.
Beyond create-and-redirect, the API supports deleting short URLs and listing recently created ones, with cache entries invalidated automatically on delete.
The API is documented automatically through FastAPI's built-in Swagger/OpenAPI support, and ships with a Docker Compose setup for local services.

Key Highlights

Hash-based short ID generation using SHA-256 + Base62 encoding, scoped per user
Three-tier caching architecture with Redis (7-day TTL, refreshed on every hit)
Full CRUD API - create, redirect, delete, and list recent short URLs
Async database operations using Motor (MongoDB async driver)
Deterministic hashing with salted re-hash collision handling
Automatic cache invalidation and TTL-based expiry
Scalable MongoDB document storage for long-term URL mapping
Horizontal scalability with a stateless, Docker-ready API design

Tech Stack

FastAPI for async request handling and automatic API docs
MongoDB for durable, horizontally scalable URL storage
Redis for sub-millisecond cache reads on hot redirects
Docker for a consistent, reproducible deployment environment

How It Works

Incoming URLs get a short ID from a SHA-256 hash of the long URL plus user ID, encoded in Base62 for a compact, URL-safe string. If that ID already exists, a UUID salt gets appended and the URL is re-hashed, which keeps the scheme deterministic without needing a central counter.

MongoDB holds the long-term URL mappings. It was picked for its flexible document model and the ability to scale horizontally as the dataset grows.

A three-tier Redis cache sits in front of MongoDB with a 7-day TTL that resets on every access, so hot redirects resolve in well under a millisecond without ever touching the database.

Request → Redis (Cache) → MongoDB (Source of Truth) ↓ miss ↓ found Query DB --------→ Cache result

Database calls run through Motor, MongoDB's async driver, so the FastAPI service can handle many concurrent redirects without blocking.

Beyond create-and-redirect, the API supports deleting short URLs and listing recently created ones, with cache entries invalidated automatically on delete.

The API is documented automatically through FastAPI's built-in Swagger/OpenAPI support, and ships with a Docker Compose setup for local services.

Key Highlights

Hash-based short ID generation using SHA-256 + Base62 encoding, scoped per user

Three-tier caching architecture with Redis (7-day TTL, refreshed on every hit)

Full CRUD API - create, redirect, delete, and list recent short URLs

Async database operations using Motor (MongoDB async driver)

Deterministic hashing with salted re-hash collision handling

Automatic cache invalidation and TTL-based expiry

Scalable MongoDB document storage for long-term URL mapping

Horizontal scalability with a stateless, Docker-ready API design