Docker

Official image: ghcr.io/arjun988/aimake

Published from the repo Dockerfile on pushes to the default branch and version tags (v*). Multi-stage build installs aimake with extras s3, huggingface, and experiments on Python 3.12-slim. Entrypoint is aimake.

This page is about running aimake in a container. To wrap your pipeline steps in images, use the Docker plugin instead.

Related: CI/CD, TypeScript SDK (aimake serve), Python SDK.

---

Pull

docker pull ghcr.io/arjun988/aimake:latest

Useful tags (from GHCR metadata): latest, semver (2.0.0, 2.0), branch names, and short git SHAs.

---

Run builds

Mount the project and set the working directory to /workspace (the image WORKDIR):

docker run --rm -v "$PWD:/workspace" -w /workspace \
  ghcr.io/arjun988/aimake:latest build

docker run --rm -v "$PWD:/workspace" -w /workspace \
  ghcr.io/arjun988/aimake:latest plan

docker run --rm -v "$PWD:/workspace" -w /workspace \
  ghcr.io/arjun988/aimake:latest doctor

Pass CLI args after the image name (entrypoint is already aimake):

docker run --rm -v "$PWD:/workspace" -w /workspace \
  ghcr.io/arjun988/aimake:latest build evaluation --jobs 4

docker run --rm -v "$PWD:/workspace" -w /workspace \
  ghcr.io/arjun988/aimake:latest --project apps/rag build

On Windows PowerShell, use ${PWD} or an absolute path for the volume mount.

---

Serve API / dashboard backend

docker run --rm -v "$PWD:/workspace" -w /workspace -p 8765:8765 \
  ghcr.io/arjun988/aimake:latest serve --host 0.0.0.0 --port 8765

Point the dashboard or @aimake/sdk at http://localhost:8765.

---

Environment and secrets

Forward tokens the same way you would locally:

docker run --rm \
  -v "$PWD:/workspace" -w /workspace \
  -e HF_TOKEN -e WANDB_API_KEY -e AWS_ACCESS_KEY_ID -e AWS_SECRET_ACCESS_KEY \
  ghcr.io/arjun988/aimake:latest build

Do not bake secrets into the image. Prefer env / your secret store — see Security.

---

GitHub Actions

Option A — docker URI action

- uses: docker://ghcr.io/arjun988/aimake:latest
  with:
    args: build

Option B — explicit docker run

- name: aimake doctor + build
  run: |
    docker run --rm -v "$PWD:/workspace" -w /workspace \
      ghcr.io/arjun988/aimake:latest doctor
    docker run --rm -v "$PWD:/workspace" -w /workspace \
      ghcr.io/arjun988/aimake:latest build

Option C — official composite Action

Prefer the first-party action when you want cache helpers and PR comments:

- uses: arjun988/aimake/.github/actions/aimake@v2
  with:
    config: aimake.yaml
    extra: s3

See CI/CD for full workflows.

---

Build the image locally

git clone https://github.com/arjun988/aimake
cd aimake
docker build -t aimake:local .
docker run --rm -v "$PWD:/workspace" -w /workspace aimake:local --help

Upstream publish workflow: .github/workflows/docker.yml → ghcr.io/<owner>/aimake.

---

Image vs Docker plugin

	GHCR image	Docker plugin
Purpose	Run aimake CLI/API	Run artifact commands inside your images
Config	docker run … aimake …	plugins.docker + metadata.docker
Docs	This page	Plugins → Docker