Welcome to the Agency repository! This project showcases the power and flexibility of the Agency Swarm framework by implementing custom AI agent teams for various problem-solving scenarios.
The Agency is a collection of specialized AI agencies, each designed to tackle specific types of problems or domains. Currently, our main focus is the Problem Solving Agency, a sophisticated AI-driven system that breaks down complex problems into manageable tasks and solves them using specialized AI agents.
The Problem Solving Agency is an advanced AI system that:
- Breaks down complex problems into structured, actionable tasks
- Generates specialized AI agent profiles for each task
- Executes tasks using tailored AI agents
- Synthesizes comprehensive solutions from individual task results
For more details, see the Problem Solving Agency README.
More agencies will be added in the future.
- Python 3.12 or higher
- Docker (recommended for containerized deployment)
-
Clone the Repository:
git clone https://github.com/husniadil/agency.git cd agency
-
Choose an Agency:
Navigate to the directory of the agency you want to run (e.g.,
problem_solving_agency
). -
Install Requirements:
pip install -r requirements.txt
-
Set Up Environment Variables:
Create a
.env
file in the agency directory and add your OpenAI API key:OPENAI_API_KEY=your_openai_api_key_here
-
Run the Agency:
python agency.py
-
Create a
stack.env
file:Create a
stack.env
file in the root of the project and add your OpenAI API key:OPENAI_API_KEY=your_openai_api_key_here
-
Build and Run with Docker Compose:
From the root of the project, run:
docker compose up
This will build and start all agencies defined in the
docker-compose.yml
file. -
Access the Agency:
The Problem Solving Agency will be available at
http://localhost:10007
(or the port specified in thedocker-compose.yml
).
agency/
├── problem_solving_agency/
│ ├── problem_solver/
│ │ ├── tools/
│ │ │ ├── agent_executor.py
│ │ │ ├── agent_profile_generator.py
│ │ │ ├── task_breakdown.py
│ │ │ └── ...
│ │ ├── problem_solver.py
│ │ └── instructions.md
│ ├── agency.py
│ ├── agency_manifesto.md
│ ├── Dockerfile
│ ├── requirements.txt
│ └── README.md
├── docker-compose.yml
├── .gitignore
└── README.md
We welcome contributions to the Agency! If you have ideas for new agencies, improvements to existing ones, or new features, please open an issue or submit a pull request.
This project is licensed under the MIT License - see the LICENSE file for details.
- This project is built using the Agency Swarm framework.
- Special thanks to the OpenAI team for providing the GPT models that power our AI agents.
Thank you for exploring the Agency. Let's push the boundaries of AI problem-solving together!