Company Profile · Data Infrastructure
Mage.
The developer-first data pipeline platform that set out to make Apache Airflow feel human again - open-source at the core, AI-assisted at the edges.
The developer-first data pipeline platform that set out to make Apache Airflow feel human again - open-source at the core, AI-assisted at the edges.
Every dashboard an executive trusts, every machine-learning model a product ships, and every report a finance team files rests on an unglamorous foundation: the data pipeline. It is the plumbing that moves raw information from where it is created to where it is used, cleaning and reshaping it along the way. When it works, no one notices. When it breaks - silently, at 2 a.m. - someone's morning is ruined. Mage, a company founded in 2020 and based in the Santa Clara area of California, exists to make that plumbing less painful to build and less likely to break.
At its simplest, Mage is a tool for building, running, and managing data pipelines that integrate and transform data. Engineers write their logic in Python, SQL, or R inside a notebook-style interface, wire the pieces together as modular blocks, and then schedule, run, and monitor the whole thing in production. The open-source project - installable with a single Docker command - has drawn a following of more than 8,800 stars on GitHub, the software world's rough measure of practitioner affection.
The company's ambition is larger than a single tool. Mage is positioned as a friendlier alternative to Apache Airflow, the workhorse orchestrator that has run data infrastructure at large companies for a decade but is widely regarded as powerful and unpleasant in equal measure. Where Airflow assumes a dedicated platform team, Mage assumes a two-person data team that needs to ship today.
"Give your data team magical powers."
That tagline is not marketing whimsy alone. The name, the wizard iconography, and the emoji that prefixes the GitHub repository all lean into the same idea: the difficult work of data engineering should feel, to the person doing it, a little bit like magic - and a lot less like a chore.
Mage's founder and chief executive, Tommy Dang, did not arrive at data infrastructure by accident. He joined Airbnb early, in 2015, and spent years building the internal developer tools, data tools, and infrastructure - including work adjacent to Airflow itself - that helped the company's engineers move quickly. He left at the end of 2020 to start Mage with co-founder Xiaoyou Wang.
The lesson he carried out the door was specific: the modern data stack had grown enormously capable but not especially pleasant. Pipelines failed without warning. Debugging meant squinting at logs. Beginners were punished by tools designed for specialists. Mage's answer was to bake the fixes into the default experience - live data previews, step-by-step execution, monitoring and alerts that come standard, and version control through git.
The pitch that first got Mage a headline was "the Stripe for AI." The product that stuck was humbler: pipelines that simply work.
When Mage surfaced in the press in 2021, it did so with an ambitious framing - developer tools to build AI into applications, "the Stripe for AI," as TechCrunch put it. The durable product turned out to be more grounded, and arguably more valuable: a data pipeline platform that engineers actually enjoyed using.
A self-hosted development environment for building modular batch and streaming pipelines in Python, SQL, and R. Notebook-style UI, prebuilt connectors to databases, APIs and cloud storage, native dbt support, and visual debugging. Install via Docker, pip, or conda.
The fully managed, scalable tier. Mage handles provisioning, upgrades, and uptime, and adds enterprise security with SSO and RBAC, team workspaces with role-based permissions, and an AI assistant that generates, debugs, and refactors pipeline code from natural language.
Block-based authoring driven by AI, plus a mage-agent command-line tool and MCP support so AI coding assistants and agents can write, fix, and monitor pipelines - Mage's step toward the "AI data engineer."
Loose scripts and manual cron jobs get replaced with monitored, scheduled runs that recover from failure instead of dying silently.
Orchestration, transformation, and monitoring live in one platform rather than a tangle of disconnected tools that never quite fit together.
Validation and standardization are built into the pipeline, so bad records get caught before they reach a dashboard or a model.
Visibility, permissions, and auditability - the governance that enterprises require and small teams rarely have time to build.
INTEGRATES WITH ›› Snowflake · BigQuery · Redshift · Amazon S3 · dbt · Tableau · Power BI · Salesforce · PostgreSQL · and more
Data orchestration is a contested market. Apache Airflow commands the widest ecosystem; Prefect leans into Python-first orchestration; Dagster centers on asset modelling; newer entrants like Kestra and Bruin keep the pressure on. Mage did not try to out-feature any of them. Its wedge was the small team that lacks a platform group and wants a friendly, block-based editor to get a pipeline into production without a multi-week ramp.
ILLUSTRATIVE POSITIONING — DEVELOPER-EXPERIENCE FOCUS
Bars reflect Mage's stated positioning on developer experience for small teams, not a benchmark of raw capability. Airflow, Dagster, and Prefect each lead on dimensions Mage does not target.
Mage has raised roughly $11.8 million across two rounds. What draws the eye is less the sum than the signatures. Gradient Ventures - Google's AI-focused venture arm - led both. Joining were Neo, Designer Fund, and Alumni Ventures, along with operator-angels who have shipped developer and consumer products themselves: Unity chief executive John Riccitiello, Behance founder Scott Belsky, "Lenny's Newsletter" author Lenny Rachitsky, and James Beshara.
| Round | Amount | Date | Lead / Notable Investors |
|---|---|---|---|
| Seed | $6.3M | Oct 2021 | Gradient Ventures, Neo, Designer Fund, + angels |
| Seed Extension | $5.5M | Mar 2023 | Gradient Ventures, Designer Fund, Alumni Ventures |
Tommy Dang leaves Airbnb at the end of the year and starts Mage with co-founder Xiaoyou Wang.
Gradient Ventures leads a $6.3M seed; TechCrunch covers Mage's developer-tools pitch.
The Apache-2.0 pipeline project builds a practitioner following as a friendly Airflow alternative.
Existing investors add roughly $5.5M, bringing total funding to $11.8M.
The managed, AI-assisted enterprise tier arrives with SSO, RBAC, and team workspaces.
AI Blocks, a mage-agent CLI, and MCP support let AI tools and agents build and monitor pipelines.
A walkthrough of Mage Pro's AI Blocks automating data analysis and transformation (2025).
▶ TUTORIALA hands-on introduction to building a pipeline end to end in the Mage interface.
▶ INTERVIEWThe founder on carving his own path at Airbnb, from data engineer to CEO of Mage.
▶ CHANNELProduction workflows across apps, APIs, warehouses, and agents on the official channel.
Mage is a data pipeline platform for building, running, and managing pipelines that integrate and transform data. Teams write pipelines in Python, SQL, and R using a notebook-style interface, then schedule and monitor them in production.
Mage (legally Mage Technologies, Inc.) was founded at the end of 2020 by Tommy Dang, a former Airbnb data-infrastructure engineer who serves as CEO, together with co-founder Xiaoyou Wang.
Yes. The core Mage project is open source under the Apache-2.0 license and can be self-hosted via Docker, pip, or conda. Mage Pro is the paid, fully managed tier that adds enterprise and AI features.
Mage offers a developer-friendly, notebook-style, block-based editor with live debugging and best practices built in, aimed at smaller teams without a dedicated platform group. Airflow is more configuration-heavy and ecosystem-broad.
About $11.8 million total - a $6.3M seed in 2021 and a roughly $5.5M seed extension in 2023 - led by Gradient Ventures, Google's AI-focused venture arm.