Neploy

Archived

One-click deployment platform that abstracts infrastructure, CI/CD, and runtime config for early-stage builders.

infra, docker, aws, node, devtools

Overview

Neploy is a developer-centric deployment platform. Push code, get a running application. No Dockerfiles to write, no CI/CD pipelines to configure, no infrastructure to provision. The platform handles build detection, containerization, and deployment to cloud infrastructure automatically.

Problem

Deploying a side project shouldn't require DevOps knowledge. But the gap between "it works on my machine" and "it's running in production" is still absurdly wide. Most early-stage builders spend more time fighting deployment than building features.

Approach

The system works in three phases:

• Detection: Analyzes the repository to identify the framework, runtime, and build requirements. Generates a build plan without requiring a config file.
• Build: Containerizes the application using optimized Dockerfiles generated from the build plan. Multi-stage builds keep images small.
• Deploy: Provisions infrastructure on AWS (ECS for containers, ALB for routing, Route53 for DNS) and deploys the container with zero-downtime rolling updates.

The entire flow is triggered by a Git push. No dashboard clicks, no config files, no manual steps.

Key Decisions

• Auto-detection over config files: Requiring a neploy.yml would defeat the purpose. The platform infers everything it can from the codebase structure and only asks for what it can't determine.
• Docker as the universal packaging format: Every application becomes a container, regardless of language or framework. This simplifies the deployment target to a single abstraction.
• AWS over building custom infra: Building a cloud from scratch would be a multi-year effort. Using AWS primitives and abstracting them behind a clean API gives 90% of the value at 10% of the cost.

Challenges

Build detection is harder than it looks. A Next.js app with a custom server, a Python ML service, and a static site all need different build strategies, but they might all live in the same repository. The detection system uses a scoring model across framework markers to determine the most likely build configuration.

Outcome

Reduced deployment time from hours to seconds for early-stage projects. The platform validated the thesis that infrastructure complexity is the biggest barrier to shipping for solo developers and small teams.

Related

Orqys — AI-native engineering ops layer that converts tickets into production-ready pull requests via multi-agent orchestration.