Loopy

Loopy is an open-source orchestrator designed for managing autonomous agent workflows that respond to changes in data. It enables the automation of multi-step processes by defining workflows as directories containing markdown files, with each file representing a step triggered by specific events or the completion of other steps. Workflows can be initiated by various sources, including webhooks, sensors, or built-in events from third-party services such as GitHub, Zendesk, Sentry, and Datadog.

The platform is agent-neutral and code-first, allowing developers to define agents, sandboxes, and typed events in a registry configuration file. Sensors, implemented as small Python functions, listen for external events—such as a webhook from Zendesk or a GitHub pull request—and translate them into typed events that are published to an event bus. Agents then subscribe to these events and execute workflow steps accordingly. For certain integrations like GitHub, Loopy provides built-in event triggers that do not require custom sensors, streamlining the process for common developer workflows such as code reviews or responding to customer feedback.

12 or newer. Users install the CLI, scaffold a new project (which generates the necessary configuration and starter workflow files), and run the engine to compile the workflow graph and start the server. The engine hosts sensors, handles webhooks, and executes agent workflows in Daytona sandboxes. The configuration supports specifying repositories to work on, injecting authentication, and managing secrets through environment files.

As an open-source tool, Loopy is positioned for developers who need a flexible, event-driven automation system that integrates with a variety of external services and supports custom agent logic. Its code-first approach and agent-neutral design aim to provide adaptability for diverse workflow automation needs.