Back to projects

PixelErase

Full-stack background removal platform with distributed asynchronous job queues.

PixelErase is a full-stack, AI-powered background removal platform engineered with Next.js, Fastify, and Python. Powered by rembg and U2-Net deep learning models, it executes high-precision background erasure via asynchronous job queues, leveraging Redis and BullMQ for reliable distributed task handling and Amazon S3/MinIO for cloud asset storage.

Tech Stack

TypeScriptNext.jsFastifyNode.jsPythonFastAPIPostgreSQLPrismaRedisBullMQAmazon S3Docker

Features

  • PixelErase is a full-stack background-removal platform built with Next.js, Fastify, and Python.
  • Uses rembg/U2-Net to remove image backgrounds through job-based API processing.
  • Integrates PostgreSQL, Redis, BullMQ, and MinIO/S3 for storage and task management.
  • Implements secure authentication, rate limiting, CORS, and remote image URL validation.
  • Supports Docker-based local development, automated testing, and production deployment.

Link