Skip to content
Klutch.sh Docs

Deploying a Mage AI App

Introduction

Mage AI is an open-source data pipeline tool for building, running, and monitoring ETL/ELT workflows. Deploying Mage AI with a Dockerfile on Klutch.sh provides reproducible builds, managed secrets, and persistent storage for projects and logs—all configured from klutch.sh/app. This guide covers installation, repository prep, a production-ready Dockerfile, deployment steps, Nixpacks overrides, sample usage, and production tips.

Prerequisites

  • A Klutch.sh account (create one)
  • A GitHub repository containing your Mage AI code/config (GitHub is the only supported git source)
  • Docker familiarity and Python 3.10+ knowledge
  • PostgreSQL credentials (if you use an external metadata DB) or plan to use the default SQLite
  • Storage for project files, logs, and data

For onboarding, see the Quick Start.

Architecture and ports

  • Mage AI serves HTTP; set the internal container port to 6789 (Mage UI default).
  • If you use PostgreSQL, deploy it as a separate Klutch.sh TCP app exposed on port 8000, connecting internally on 5432.
  • Persistent storage is recommended for the Mage project folder and logs.

Repository layout

mage-ai/
├── Dockerfile              # Must be at repo root for auto-detection
├── requirements.txt
├── start.sh                # Optional helper
├── mage_data/              # Project data/logs (mount as volume)
└── .env.example            # Template only; no secrets

Keep secrets out of Git; store them in Klutch.sh environment variables.

Installation (local) and starter commands

Install dependencies and run locally before pushing to GitHub:

Terminal window
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
mage start mage_ai

Optional start.sh for portability and Nixpacks fallback:

#!/usr/bin/env bash
set -euo pipefail
exec mage start mage_ai --port 6789 --host 0.0.0.0

Make it executable with chmod +x start.sh.

Dockerfile for Mage AI (production-ready)

Place this Dockerfile at the repo root; Klutch.sh auto-detects it (no Docker selection in the UI):

FROM python:3.11-slim


WORKDIR /app


RUN apt-get update && apt-get install -y build-essential git && rm -rf /var/lib/apt/lists/*


COPY requirements.txt /app/requirements.txt
RUN pip install --no-cache-dir -r requirements.txt


COPY . /app


ENV PORT=6789


EXPOSE 6789
CMD ["mage", "start", "mage_ai", "--port", "6789", "--host", "0.0.0.0"]

Notes:

  • Add any system dependencies your pipelines require (e.g., libpq-dev for Postgres, ffmpeg for media processing).
  • Keep mage_data/ writable and mount it as a volume for pipeline state and logs.

Environment variables (Klutch.sh)

Set these in the Klutch.sh app settings (Secrets tab) before deploying:

  • PORT=6789
  • MAGE_DATABASE_URL=postgres://<user>:<password>@<host>:<port>/<db> (if using Postgres)
  • MAGE_SECRET_KEY=<secure-key>
  • Any pipeline-specific credentials (AWS/GCP keys, API tokens)

If you deploy without the Dockerfile and need Nixpacks overrides:

  • NIXPACKS_BUILD_CMD=pip install -r requirements.txt
  • NIXPACKS_START_CMD=mage start mage_ai --port 6789 --host 0.0.0.0
  • NIXPACKS_PYTHON_VERSION=3.11

These keep Mage AI compatible with Nixpacks defaults when a Dockerfile is absent.

Attach persistent volumes

In Klutch.sh storage settings, add mount paths and sizes (no names required):

  • /app/mage_data — projects, logs, and pipeline state.
  • /app/.cache — optional cache if your pipelines write there.

Ensure these paths are writable inside the container.

Deploy Mage AI on Klutch.sh (Dockerfile workflow)

  1. Push your repository (with the Dockerfile at the root) to GitHub.
  2. Open klutch.sh/app, create a project, and add an app.
  1. Connect the GitHub repository; Klutch.sh automatically detects the Dockerfile.
  2. Choose HTTP traffic for Mage AI.
  3. Set the internal port to 6789.
  4. Add the environment variables above (database URL if used, secret key, pipeline creds, and any NIXPACKS_* overrides if you temporarily deploy without the Dockerfile).
  5. Attach persistent volumes for /app/mage_data (and /app/.cache if used), selecting sizes that match your data and logging needs.
  6. Deploy. Your Mage AI UI will be reachable at https://example-app.klutch.sh; attach a custom domain if desired.

Sample API usage

Trigger a pipeline run via HTTP (replace placeholders):

Terminal window
curl -X POST "https://example-app.klutch.sh/api/pipelines/<pipeline_uuid>/trigger" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <api-token>" \
  -d '{"pipeline_run": {"variables": {"example_var": "hello"}}}'

Health checks and production tips

  • Add a reverse proxy probe to / or a lightweight health route if available.
  • Enforce HTTPS at the edge; forward HTTP to port 6789 internally.
  • Keep dependencies pinned in requirements.txt and test upgrades.
  • Monitor disk usage on /app/mage_data and resize volumes before they fill.
  • Back up PostgreSQL (if used) and project data regularly; do not rely on container storage alone.
  • Rotate secrets and API tokens regularly; store them only in Klutch.sh secrets.

Mage AI on Klutch.sh combines reproducible Docker builds with managed secrets, persistent storage for pipeline state, and flexible HTTP/TCP routing. With the Dockerfile at the repo root and port 6789 configured, you can run reliable data pipelines without extra YAML or workflow overhead.