🌱 Welcome to Airflow Copilot
A conversational AI assistant that lets you run, monitor, and manage Apache Airflow directly from Microsoft Teams — powered by LLMs and the Airflow REST API.
✨ What is Airflow Copilot?
Airflow Copilot bridges the gap between orchestration and conversation.
It enables teams to:
- 🗣️ Interact with Airflow using natural language inside Microsoft Teams
- 🔁 Trigger and monitor DAGs, tasks, schedules
- 🧠 Leverage LLMs (OpenAI, Google, Anthropic, Groq)
- 🔒 Enforce RBAC-based access control
- 📚 Preserve memory via LangGraph checkpoints in PostgreSQL
🚦 Ready to Begin?
💡 Follow the Quick Start and get your Copilot up & running in less than 10 minutes.
| 🔹 Section | 🔗 Link | 📌 Why Visit? |
|---|---|---|
| 🚀 Quick Start | Get Started → | Two-file Docker setup for local testing |
| 🛠️ Architecture | How it Works → | High-level design & component workflow |
| ⚙️ Configuration | Env Variables → | All runtime settings at a glance |
| 📦 Deployment | Deploy Options → | Local, Docker Prod, Kubernetes, AWS |
| 🤖 Agent Behavior | Agent Logic → | Fallbacks, retries, limitations |
🧰 Features Overview
- 🔄 Chat-driven orchestration of DAGs, variables, tasks, and runs
- 🧠 LLM-powered intelligence via OpenAI, Gemini, Claude, Groq
- 🔐 Role-aware execution with Airflow RBAC integration
- 🧩 Persistent memory stored via LangGraph + PostgreSQL
- 🌍 Unified deployment: Docker Compose → Kubernetes → ECS
🧪 Try It Out
🛠️ Follow the Quick Start → chat with your DAGs in Microsoft Teams.
🤝 Let’s Build Together
Airflow Copilot is fully open source.
We welcome your ideas, tools, feedback, and pull requests.
👉 View the Roadmap · Contribute
🚀 Automate smarter. Collaborate better. That’s Copilot.