Skip the complex configs. pooolify is the open-source framework for building powerful AI agent teams with just an intuitive folder structure and pure Python.
No reviews yetBe the first to leave a review for pooolifyAI
Maker
📌
Hey Product Hunt community! 👋
We're the creators of pooolifyAI. Thanks so much for checking out our project.
Like many of you, we've been fascinated by the power of multi-agent AI systems. But as we started building with existing frameworks, we grew frustrated. We found ourselves wrestling with complex YAML files, debugging opaque "black boxes," and struggling to manage the chaotic interactions between agents. It felt like we were losing control instead of gaining a powerful new capability.
That's why we built pooolifyAI with a different philosophy.
Instead of pre-defining every possible step, pooolifyAI empowers a central Manager AI to plan, delegate, and intelligently orchestrate a team of specialist agents.
Everything is defined in pure, simple Python, because as developers, that’s where we’re most productive. We designed an intuitive project structure: new agent? Just create a new folder. It’s that simple. This code-native approach means no more context-switching or fighting with syntax.
With pooolifyAI, you get:
✨ Frictionless Developer Experience: Get started in minutes with an intuitive, folder-based structure. We even include a pure HTML/JS test UI right out of the box, so you can start interacting with your agents immediately—no frontend setup needed.
🤖 A Manager-led Architecture: Forget designing complex flows. A central Manager AI automatically plans the entire workflow, deciding which agents to run in parallel and how data flows between them. You just define the team and the goal.
🔍 Full Process Transparency: Our API is asynchronous. While your request is processed, you can poll for updates and get a detailed, structured log of every step. You'll see the full thought process, plans, and tool results.
🔌 Model Agnosticism: Our vision is to support any model. For this initial launch, we're optimized for GPT-5 and GPT-5-high. Full support for Gemini, Claude, and open-source models is at the top of our roadmap!
This is an early-stage, open-source project, and we're truly here to learn from this amazing community. Our goal is to build the most developer-friendly framework for creating powerful and transparent AI agent systems.
We'd be thrilled to get your feedback:
What kind of agent teams would you want to build with this?
Does the "Manager-led" architecture resonate with the problems you're trying to solve?
What's the one feature you'd love to see next?
You can check out the project on GitHub, and thanks to the built-in web UI and simple structure, you can start testing it in minutes. We've tried to make it as frictionless as possible to get started. If you find it interesting, a star ⭐ would mean the world to us!
We'll be here all day to answer every single question. Let's build the future of AI collaboration together!