added a folder (babyagi-ui) with a UI to use the system

[Nkluge-correa]

Hello hello 👋

I created a simple UI for the TDAA (Task-Driven Autonomous Agent, a.k.a babyagi). All credits for the code implementation goes to Yohei Nakajima, how made a clean, simple, and easy to replicate implementation. This UI enables users to have an "easier" interface to experiment with the TDAA.

In this version, the controller can set how many rounds the TDAA agent will optimize its tasks. Also, the memory deque list was set so that the agent can optimize more than one objective in a single session, i.e., after completing the specified round, you can set a new objective/task, and the previous task won't affect the new rounds. If you wish to extend the memory of the TDAA, just add more rounds.

Controls ⚙️

• Rounds: Number of times the TDAA repeats an execution loop (completing the task, generating new tasks, prioritizing tasks).
• Objective: The goal you are trying to attain.
• First Task: An initial task to get things started.

Requirements 🛠️

dash
openai
pinecone-client
dash-bootstrap-components

Set your API keys in the config.env file, and have fun!

[tnachen]

This is cool! any screenshot to share?

[Nkluge-correa]

For sure, is nothing much, but it helps people that dont like to work from the CLI. Dash apps are really easy to build.