We plan to refactor into a more capable agent soon! It's exciting to know that 2 years before the recent Anthropic paper our team was showing the power of code execution as a tool through experiments like UnrealGPT. With all the recent progress since this was first released will be exciting to see how far it can be pushed with SoTA models and agentic scaffolding.
Next version should have much better long horizon task planning, multi-modal input (it can see what it's building), and possibly multi-agent hand-offs.
Cheers,
Your friends at TREE
We plan to refactor into a more capable agent soon! It's exciting to know that 2 years before the recent Anthropic paper our team was showing the power of code execution as a tool through experiments like UnrealGPT. With all the recent progress since this was first released will be exciting to see how far it can be pushed with SoTA models and agentic scaffolding.
Next version should have much better long horizon task planning, multi-modal input (it can see what it's building), and possibly multi-agent hand-offs.
Cheers,
Your friends at TREE