OVADARE

I’ve been working with AutoGen for a while now and kept running into a challenge—AI agents don’t always stay in sync. Unlike humans, they don’t share social norms, priorities, or an inherent way to resolve conflicts when goals misalign. That’s why I built OVADARE, an open-source framework designed to detect, resolve, and learn from conflicts between AI agents in multi-agent platforms like AutoGen and CrewAI. It runs alongside these frameworks, helping agents stay aligned, avoid redundant work, and prevent decision loops that disrupt workflows. Haven't officially lunched yet but the founders of CrewAI and AutoGen had already reached out, which was really exciting. Now, I’d love to get more thoughts from the this awesome community. If you’ve ever had agents work at cross-purposes or break a workflow, give OVADARE a try and let me know what you think.