Why Agentree

AI agents have become extremely capable. But, UI wise, we are still stuck with chat interfaces. Chat interfaces have two fundamental limitations -

Linear Thinking

Chat interfaces have fundamental limitation that it restricts you to linear thinking. Real world software development is messy - you want to test out multiple hypotheses, ask stupid questions, multiple ideas emerges from one. Chat interfaces force you to either limit yourself to linear work or poisson your context.

Lack Of High Level Overview

If you use agents a lot, you'd have faced this problem with long conversations that it becomes difficult to keep track of what was discussed. Chats always show you very detailed zoomed in version. In Agentree, you'll always have bird's eye view of entire conversation with ability to watch any message by just hovering on it.

Collaboration

• Well, you can't collaborate on a chat .. which is fundamental limitation of being linear.

When most of the code is written by AI agents, it's important not to just share code with your team .. but the process of how that code came into existence - the process, the decisions user made, most importantly - the things that were tried but did not work out, other possibilities that should be tried etc.

Agentree

Agentree's main promise is providing proper version control for Agent conversations. And, that unlocks many downstream possibilities - better team collaboration, better monitoring of parallel agents, providing precise context to agents, and even better collab between running parallel agents.

In agentree, all messages are saved in graph as nodes and all agent code changes are persisted as git commits. This allows you to branch from any point of conversation with full confidence that you can easily retun later.

Dev's personal experience - Agentree makes running and monitoring multiple agents so much easier that I end up spending 10x more tokens with less hair pulling.