Mage AI

raw JSON →
0.9.79 verified Sat May 09 auth: no python

Mage is a modern data pipeline tool for building, running, and managing data pipelines. It integrates with Python, SQL, and R, and provides a web-based UI for development and monitoring. Current version: 0.9.79. Release cadence: frequent (multiple releases per month).

pip install mage-ai
error ModuleNotFoundError: No module named 'mage_ai'
cause Mage is not installed or installed in a different environment.
fix
Run pip install mage-ai and ensure you're using the same Python environment (e.g., virtualenv, conda).
error ImportError: cannot import name 'start' from 'mage_ai.server.app'
cause The import path changed; older versions had a different structure.
fix
Upgrade to latest using pip install --upgrade mage-ai and use the correct import: from mage_ai.server.app import start.
error AttributeError: module 'mage_ai' has no attribute 'init'
cause Trying to run `mage init` as a Python function instead of a CLI command.
fix
Run mage init project_name in your terminal, not in Python.
gotcha Mage's default server is designed for development; for production, use a production-grade ASGI server like uvicorn or gunicorn.
fix Use `uvicorn mage_ai.server.app:app --host 0.0.0.0 --port 6789` instead of calling start() in production.
gotcha The `init` command is not available inside Python code; it must be run from the CLI.
fix Run `mage init my_project` from the terminal, not from Python.
deprecated Older Mage versions used `mage_ai.orchestration` for pipeline execution; this has been restructured.
fix Upgrade to latest and use `mage_ai.data_preparation` for pipeline definitions.

Start the Mage development server programmatically.

import os
from mage_ai.server.app import start

# Set environment variables for authentication (optional)
os.environ.get('MAGE_API_KEY', '')

# Start Mage server
start()