Dagu

•
4.6k
•

Orchestrates lightweight, powerful workflows in a single binary with a Web UI across Linux, macOS, and Windows, enabling AI agents to create, update, and run them with full traceability.

Dagu is a simple, lightweight, and local-first workflow engine packaged as a single binary, ideal for small teams and private infrastructure across Linux, macOS, and Windows. It streamlines operational automation and data script management by providing scheduling, dependencies, retries, and logging without requiring external databases or brokers. Uniquely, Dagu integrates with AI agents via its built-in MCP server, allowing intelligent automation for creating, improving, debugging, and executing workflows, all while offering a comprehensive Web UI for observability and control.

Key Features

01Local-first, file-based workflows with a single binary and no external database required
02Durable execution with scheduling, dependencies, retries, queues, logs, and notifications
03Integrated AI agent support via a built-in MCP server for intelligent workflow creation and execution
043,418 GitHub stars
05Language-agnostic execution for shell scripts, containers, SQL queries, and HTTP requests
06Built-in secret management, human-in-the-loop approvals, and RBAC with SSO support

Use Cases

01ETL and data operations, turning scripts and queries into observable pipelines
02Cron and legacy script management with a UI, logging, retries, and notifications
03Infrastructure and server automation, running commands or scripts over SSH on remote servers