What is Ordine?

🚧 Preview

Ordine is currently in Preview stage. APIs and features may change. We welcome feedback and contributions!

Ordine is an AI Agent first work orchestration framework. It provides a universal work orchestration framework for defining, composing, and executing automated workflows — powered by AI agents and scripts.

Code quality automation is Ordine's first and flagship use case, delivered as a built-in plugin. Over time, domain-specific capabilities (code review, security scanning, documentation generation, etc.) will be fully extracted into plugins, leaving Ordine as a pure orchestration core.

The Problem

Automation workflows are everywhere — code review, CI/CD, data processing, content generation — yet each domain reinvents its own orchestration layer. Teams end up with fragmented tools that can't compose, share context, or leverage AI agents effectively.

The Solution

Ordine provides a single orchestration layer that:

1. Define operations — atomic tasks executed by AI agents, scripts, or plugins
2. Compose operations into multi-step pipelines (DAG execution)
3. Extend capabilities through a plugin system
4. Automate execution through rules that trigger pipelines on events

Code Quality Plugin (Built-in)

The built-in code quality plugin adds:

• Best Practices — machine-readable coding standards with checklists
• Code-aware Operations — check/fix code using AI agents or linters
• Rule Triggers — automatically run pipelines on code changes

This plugin demonstrates Ordine's extensibility model: domain knowledge lives in plugins, orchestration lives in the core.

Key Differentiators

AI Agent First Design

Every feature is designed so that AI agents can discover, invoke, and compose it with minimal friction. Operations can use Claude, Codex, or custom scripts as backends.

Declarative Configuration

Pipelines, operations, and skills are data-driven. Define them as JSON/YAML, not imperative code. This makes them portable, version-controllable, and agent-accessible.

Typed Pipeline Engine

The pipeline engine uses a directed acyclic graph (DAG) with typed nodes and edges. Each node has well-defined inputs and outputs, enabling safe composition.

Extensible Architecture

• Plugins — encapsulate domain-specific knowledge and capabilities
• Skills — plug in new AI capabilities
• Operations — define custom tasks with any executor backend
• Node Types — folder, code-file, operation, output, compound, condition, github-project

Plugin System

Domain logic is progressively migrating into plugins. The core engine provides:

• Pipeline DAG orchestration
• Agent dispatching
• Entity management (operations, skills, rules, jobs)
• UI canvas and API

Plugins provide domain-specific entities, operations, and UI extensions. The code quality plugin is the first — more will follow.