dagger_gleam

Note: This package is in early development. The API may change before 1.0.

Type-safe Dagger pipelines in Gleam.

Dagger lets you define CI/CD pipelines as code, running them in containers with full caching and reproducibility. dagger_gleam brings that to Gleam with a generated, type-safe DSL and a functional execution model: you build a description of your pipeline first, then execute it — no hidden side effects.

Installation

gleam add dagger_gleam

The DSL modules (dagger/dsl/container, dagger/dsl/git, …) are pre-generated from Dagger’s GraphQL schema and included in the package. No code generation step required for users.

Quick Start

import dagger
import dagger/dsl/container as c
import dagger/dsl/host as h
import gleam/io
import gleam/string

pub fn main() {
  use client <- dagger.connect()

  let src = h.host() |> h.directory(".", with: h.none)

  let pipeline =
    c.container(with: fn(o) { o |> c.opt_platform("linux/amd64") })
    |> c.from("ghcr.io/gleam-lang/gleam:v1.14.0-erlang")
    |> c.with_mounted_directory("/src", src, with: c.none)
    |> c.with_workdir("/src", with: c.none)
    |> c.with_exec(["gleam", "test"], with: c.none)

  use result <- c.stdout(pipeline, client)
  case result {
    Ok(out) -> io.println(string.trim(out))
    Error(e) -> io.println_error(string.inspect(e))
  }
}

Run it with:

dagger run gleam run

The with: API

Every DSL function that accepts optional arguments takes a with: builder:

// no options
c.with_exec(["gleam", "test"], with: c.none)

// one option
c.with_exec(["gleam", "test"], with: fn(o) { o |> c.expand(True) })

// multiple options
c.with_exec(["gleam", "test"], with: fn(o) {
  o |> c.expand(True) |> c.skip_entrypoint(True)
})

c.none is just the identity function — use it whenever you have nothing to configure.

Examples

The examples/ directory in the source repository contains runnable examples:

How It Works

dagger.connect opens a connection to the Dagger engine and passes a Client to your callback. DSL calls build a lazy operation tree — nothing is sent over the wire until you call a terminal function (stdout, export, entries, …). The interpreter then resolves any deferred values (e.g. local directory IDs) and executes a single GraphQL query.

Source & Contributing

github.com/avit-io/dagger_gleam

Issues and PRs are welcome. Active development happens on the dev branch.