What Is CI/CD? Continuous Integration & Delivery

Quick answer
Continuous integration (CI) builds and tests every code change automatically, so bugs surface in minutes instead of weeks.
Continuous delivery (CD) keeps every passing build ready to deploy with a single click.
Continuous deployment takes CD one step further and ships every passing change to production with no manual step.

What is CI/CD?

CI/CD is a method of shipping software that automates the steps between writing code and running it for users. Continuous integration merges each change into a shared branch and immediately builds and tests it. Continuous delivery or deployment then automates pushing the result toward production. A tool that runs these steps in sequence is called a pipeline.

How does a CI/CD pipeline work?

A pipeline is a sequence of jobs triggered by an event, usually a push or pull request. A typical pipeline runs these stages in order: check out the code, install dependencies, run linting and tests, build an artifact (a deployable package or container image), and deploy it. If any stage fails, the pipeline stops and reports the failure back to the developer. Pipelines are defined as code, commonly in a file like .github/workflows/ci.yml, so the process is reviewable and reproducible.

Why does CI/CD matter?

CI/CD replaces slow, error-prone manual releases with a repeatable process. It catches integration bugs early, makes deployments boring and predictable, and lets a team ship several times a day instead of once a quarter. It is also the foundation other practices build on: security scanning, GitOps, and feature flags all run as stages inside a pipeline.

Where this bites vibecoders
An AI coding assistant can produce a working app in an afternoon, but it will not set up a pipeline unless you ask. The result is a codebase that only ever ran on one person’s laptop, with no tests and no reproducible build. The first “it works on my machine but not in production” moment usually happens here. Adding CI after the fact means retrofitting structure onto code that was never built to be tested automatically.

Where AI coding assistants get this wrong

Generating a .github/workflows file that references actions or versions that don’t exist, so the first pipeline run fails on YAML syntax.
Hardcoding secrets like API keys directly in the workflow file instead of using the platform’s secrets store.
Writing a pipeline with no test stage at all, so CI becomes “it compiled” rather than “it works.”
Ignoring the difference between delivery and deployment and configuring automatic production deploys the owner never asked for.

Checklist

Put the pipeline definition in version control, in the same repo it builds.
Make every push trigger a build, and every pull request trigger tests.
Store all secrets in the platform’s secret store, never in the YAML.
Add at least one failing-test scenario to confirm failures actually stop the pipeline.
Document what “deploy to production” means and who is allowed to trigger it.

FAQ

What is the difference between CI and CD?

Continuous integration is the automated build-and-test of every change. Continuous delivery makes passing builds releasable with a manual approval, while continuous deployment releases them automatically. The “CD” in CI/CD usually means delivery, with deployment as an optional stronger step.

Do I need CI/CD for a solo project?

Yes. CI/CD is cheap for a single developer and catches problems that only appear on a clean machine: missing dependencies, untracked files, and environment differences. For a solo or AI-assisted project it acts as a safety net that runs tests you might otherwise forget.

What tools run CI/CD pipelines?

GitHub Actions, GitLab CI/CD, and CircleCI are common hosted options, while Jenkins runs self-hosted pipelines. For beginners, GitHub Actions is usually the easiest because it lives in the repository and has a large library of ready-made actions. See How to Set Up a CI/CD Pipeline With GitHub Actions.

Is CI/CD the same as DevOps?

No. CI/CD is a specific automation practice. DevOps is the broader culture and set of practices for running software end to end, of which CI/CD is one part. See What Is GitOps? for an adjacent automation practice.