Initial problem
I wanted to use different AI models in one place, switch between providers, and organize chats around my own projects. That led me to build a workspace with a consistent interface and reliable streaming, including after a connection drops.


