Prefect is acquiring Dagster while keeping both projects independent and open source, creating a unified AI orchestration stack built around Prefect, Dagster and FastMCP without disrupting existing users.
Prefect has agreed to acquire Dagster Labs, bringing together two of the most widely adopted open-source successors to Apache Airflow while pledging to keep both projects independent and actively maintained. The combined portfolio will include Prefect for workflow orchestration, Dagster for asset-based declarative orchestration and FastMCP, the open-source implementation of the Model Context Protocol (MCP) for AI agents.
The company said Dagster will retain its name, open-source licence and roadmap, with existing engineering teams continuing development. Both open-source Dagster and Prefect will continue receiving maintenance releases, new features and security updates. Existing users are not required to migrate and can continue using either platform or deploy both together.
Prefect said the merger combines runtime workflow execution, declarative asset-based orchestration and AI agent governance to support production AI, machine learning and data engineering workloads. Future engineering efforts will focus on full-stack agentic automation, improved reliability through upstream orchestration and combining flexible execution with declarative workflow modelling.
FastMCP remains central to the strategy. According to Prefect, the project recorded more than 92 million downloads in the past month and over 26,000 GitHub stars. The company also said Anthropic adopted FastMCP as the official MCP SDK and that the two organisations continue collaborating on the Python SDK and developer experience.
Dagster founder Nick Schrock and CEO Pete Hunt will serve as strategic advisers while remaining active in the open-source community. Prefect also said its profitability over the past year provides a sustainable long-term foundation for the stewardship and continued growth of the open-source projects.
