🚀 Meet Divan: My New Open-Source AI Project What if you could organize a team of AI agents to help you build software products? Not just one AI that writes code. An entire team with different responsibilities. That’s the idea behind Divan, a new open-source project I’ve started working on. Today, AI coding tools are becoming incredibly powerful. They can help developers build websites, mobile applications, and complex software much faster. But building software isn’t only about writing code. You also need to understand the problem, define requirements, design the architecture, test the solution, check security, and prepare releases. In a traditional software company, these responsibilities are usually divided among different specialists. Divan brings a similar structure to AI agents. 🤖 Here’s how it works: 📋 A Product Agent defines what needs to be built. 🏗️ An Architecture Agent plans the technical solution. 💻 A Development Agent writes the code. 🔍 A Review Agent checks the implementation. 🛡️ A Security Agent evaluates potential risks. 📦 A Release Agent prepares the product for delivery. And throughout the process, a human remains responsible for the important decisions. The idea is simple: AI agents execute. Humans decide. The project is still in its early experimental stage, and I’m testing the framework on real software development workflows. I’ll also be sharing the challenges, failures, improvements, and lessons learned. Divan is completely open-source, and developers are welcome to explore, test, and contribute. 🔗 GitHub: github.com/farjadp/Divan I believe the next major step in AI-assisted software development is not simply making AI write code faster. It’s learning how to organize AI systems to work together more intelligently. That’s what I’m exploring with Divan. #Divan #OpenSource #AIAgents #AIDevelopment #SoftwareEngineering #AgenticAI #ArtificialIntelligence