Getting Started with Dagger: Installation, Example CI/CD Pipeline, and CUE‑Based Pipeline Definition

Dagger is a new DevOps platform announced by Docker founder Solomon Hykes, designed to let developers describe CI/CD pipelines declaratively using the CUE language.

Installation : on macOS with Homebrew run brew install dagger/tap/dagger ; on other systems use the curl script curl -L https://dl.dagger.io/dagger/install.sh | DAGGER_VERSION=0.2.4 sh . After installation, verify the binary location with type dagger and ensure Docker Engine is running because Dagger executes tasks inside Docker.

Example pipeline : the official todo sample demonstrates a full CI/CD flow. Clone the repository, checkout a version, then run dagger do build . The first run installs dependencies and takes longer; subsequent runs are faster thanks to caching.

The pipeline builds the app, writes the output to a _build directory, and can be previewed in a browser with open _build/index.html . Modifying source files (e.g., src/components/Form.js ) and re‑running the build shows the updated result instantly.

Pipeline definition : Dagger pipelines are written in CUE. A typical #Plan includes a client section for filesystem, environment, and network interactions, and an actions section defining steps such as deps , test , build , and deploy . The CUE example in the article shows how to mount caches, run Bash scripts, and use imported packages like universe.dagger.io/alpine and universe.dagger.io/docker .

Key client interactions include reading/writing files ( client.filesystem ), accessing sockets ( client.network ), handling environment variables ( client.env ), executing commands ( client.commands ), and retrieving platform information ( client.platform ).

Building container images : Dagger can build images directly from CUE without a Dockerfile. One approach uses docker.#Dockerfile with an inline Dockerfile definition; another uses docker.#Build with explicit steps ( docker.#Pull , docker.#Copy , docker.#Run , docker.#Set ) to replicate a Dockerfile declaratively.

In summary, Dagger leverages CUE to provide a portable, code‑first CI/CD toolkit. While it requires learning CUE, once mastered the pipeline configuration becomes straightforward, offering fast local testing, caching, and the ability to run pipelines anywhere.