Mage

Mage is an open-source data pipeline and ETL tool with a managed cloud edition, Mage Pro, for data engineers who build, schedule and repair pipelines in code.

Mage runs the code that moves and shapes a company's data, from loading a source to rebuilding tables in the warehouse. Mage Technologies, Inc. of Santa Clara, California, builds it in two editions. Mage OSS is open source under the Apache 2.0 licence and runs on a team's own machines with Docker. Mage Pro is the managed edition, and a project moves across with zero code changes. Work in Mage is organised as pipelines made of blocks, and each block is its own file in the project. Data integration pipelines replicate SaaS, database and file sources, with incremental replication on bookmark columns and change data capture for PostgreSQL. Transformation runs in Python, SQL or dbt blocks, so dbt models are built and scheduled beside the loads. Schedule triggers take cron expressions, sensor blocks wait on a condition, and dynamic blocks fan out at runtime. On top sit the AI features: building a pipeline from a description, creating charts and self-healing pipelines that recover failed runs. Mage suits data engineers and analytics engineers who write their pipelines as code and want loading, transformation and monitoring in one workspace. Solo builders start on Mage OSS and move to Mage Pro once they would rather not run the infrastructure themselves. Teams with shared production workflows use Mage Pro's controls and deployment workflows. Organisations with strict data rules choose private or hybrid deployment, where the data plane stays in their own AWS, GCP or Azure account. Companies leaving older ETL tools such as Informatica can have Mage rebuild their workflows on the platform. A common Mage pipeline pulls new Stripe customers, matches them to Salesforce accounts and writes clean records to BigQuery every hour. Others replicate a PostgreSQL database into Snowflake, write Iceberg or Delta Lake tables to Amazon S3, or run dbt models once a load lands. Workers run on Kubernetes, ECS, Cloud Run or Azure Container Instances in the team's own cloud. Backfills re-run a pipeline over a past date window after a fix. When a run fails, the alert reaches Slack, Microsoft Teams or Opsgenie, and the block run logs show which step broke.