[Docker](https://www.docker.com)

Docker is a platform that packages an application together with everything it needs to run, code, libraries, system tools, and configuration, into a single unit called a container. Instead of installing dependencies directly on a server and hoping they match what worked on your laptop, you build a container once and run it anywhere Docker is installed, from a developer’s machine to a production cluster. For developers building production systems, Docker matters because it removes the “it works on my machine” problem, standardizes how teams ship code, and underpins nearly every modern CI/CD pipeline, whether the deployment target is a single laptop or a fleet of production servers.

How Docker works

Docker uses a client-server setup. You run commands through the Docker client, and a background process called the Docker daemon does the actual work of building, running, and distributing containers.

Docker image: A Docker image is a read-only template that bundles your application code with its dependencies. You build an image once from a Dockerfile, a text file listing the steps to assemble it, and reuse that image everywhere.

Container: A container is a running instance of an image. It behaves like a lightweight, isolated version of the operating system, sharing the host machine’s kernel instead of running a full guest OS the way a traditional virtual machine does.

Docker Hub: Docker Hub is a public registry where you can pull pre-built images for databases, languages, and tools, or push your own images so teammates and servers can pull them.

Docker Compose: Compose lets you define multi-container applications, for example a web server, a database, and a cache, in one YAML file and start them together with a single command.

Isolation: Each container gets its own filesystem, network interface, and process space, so you can run apps with conflicting dependencies on the same host without them interfering with each other.

What you can build with Docker

Consistent dev environment: Package your app and its dependencies into an image so every teammate, and every CI server, runs the exact same environment instead of chasing “works on my machine” bugs.

Microservices architecture: Split a large application into smaller, independently deployable services, each running in its own container, so teams can update and scale services without touching the rest of the system.

CI/CD pipeline: Use Docker to build a test environment that mirrors production exactly, so a build that passes tests in a container behaves the same way when it ships.

Local database sandbox: Spin up an isolated PostgreSQL, MongoDB, or Redis instance in seconds for local development, then tear it down without leaving anything installed on your machine.

Multi-cloud deployment: Build a container once and deploy the same image to AWS, Google Cloud, or Azure, since Docker containers run the same way regardless of the underlying cloud provider.

AI agent sandboxing: Use Docker Sandboxes to run coding agents like Claude Code in an isolated, network-restricted container, so an autonomous agent can read and edit files without full access to your host machine.