agentcloud
Welcome to agentcloud
repository! This project comprises two main components:
- Agent Backend: A Python application running autogen, communicating through socket.io.
- Webapp: A UI built using next.js, tailwind, and an express custom server.
Getting Started
To run this project up locally, you'll need docker-compose. Follow the steps below to get started:
- Environment Variables: Before starting the services, ensure you fill out any necessary environment variables in the
docker-compose.yml
file. Make sure to enter your Open AI API Secret in theOPENAI_API_KEY
field. - Start Services: Simply run the following command:
docker-compose up
This will start both the agent backend and the webapp.
Current Workflow
- A task is passed to the
team-generation
team - Based on the task, the
team-generation
team creates the ideal team to undertake the task (so, we are utilising Autogen to formulate the team) - The user can give feedback to Autogen to augment the team if they wish
- Once the user is happy with the team, they type exit (it's a bit unintuitive, we know; we're working on changing it), which ends the team generation step and immediately initiates
task-execution
using the newly formulated team, the task is then undertaken using standard Autogen workflow
License
This project is licensed under the Affero General Public License, version 3 only. Please refer to the LICENSE file for more details.
Contributions & Feedback
If you wish to contribute or provide feedback, please follow the contribution guidelines in CONTRIBUTING.md.
Thank you for exploring agentcloud
!