Chat branching is kind of like time traveling. "What if I had asked GPT4o instead of Gemini 2.5 Pro?" In Hatz, you can go back and branch into a new chat with a new LLM.
Let's say I start a chat and get my output:
But I want to see how another LLM would have answered. Instead of starting a new chat and having to switch back and forth, I can click on the edit button and change the model:
I can send the chat input to the new LLM, and even change the prompt if I wanted to!
Once you've started to branch chats, you can navigate between them easily by clicking the arrows.
