Tool Recommender

Welcome to the Tool Recommender project, created with crewAI. This project leverages a multi-agent AI system to recommend tools for specific use cases across various industries. By utilizing the powerful and flexible framework provided by crewAI, we enable our AI agents to collaborate effectively on complex tasks, maximizing their collective intelligence and capabilities.

Installation

Ensure you have Python >=3.10 <=3.13 installed on your system. This project uses Poetry for dependency management. All commands to be executed within crew folder.

1. Install Poetry if you haven't already:

pip install poetry

2. Navigate to your project directory and install the dependencies:

poetry lock
poetry install

Configuration

1. Add your keys

   • OPENAI_API_KEY to the .env file in the crew directory.
   • OPENAI_ORGANIZATION_ID to the .env file in the crew directory.
   • SERPER_API_KEY to the .env file in the crew directory.
   • AGENTOPS_API_KEY to the .env file in the crew directory. // remove agentops.init() references from main.py if not needed

2. Customize the project:

   • Modify src/crew/config/agents.yaml to define your agents
   • Modify src/crew/config/tasks.yaml to define your tasks
   • Adjust src/crew/crew.py to add your own logic, tools, and specific arguments
   • Update src/crew/main.py to add custom inputs for your agents and tasks

Running the Project

To start your crew of AI agents and begin the tool recommendation process, run this command from the root folder of your project:

poetry run crew

This command initializes the Tool Recommender Crew, assembling the agents and assigning them tasks as defined in your configuration.

Using the Tool Recommender API

The project includes a Flask-based API service. To run the service:

poetry run serve

This will start the Flask server, allowing you to interact with the Tool Recommender through HTTP requests.

Making a Recommendation Request

To get a tool recommendation, you can use the following curl command:

curl -X POST -H "Content-Type: application/json" -d '{"industry": "construction industry", "use_case": "Keep track of the deal pipeline. Let other input and view their deals. I have a 6 people team. I want to view the whole pipeline but want to control what others can see.", "other_requirements": "It should be easy to use. Should be cost effective, yet scalable when I grow the company. Permissions and access controls are important."}' http://127.0.0.1:5000/run_crew

This will trigger the AI agents to analyze your requirements and provide a recommendation.

Understanding Your Crew

The Tool Recommender Crew consists of multiple AI agents, each with unique roles, goals, and tools. These agents collaborate on a series of tasks to provide comprehensive tool recommendations. The crew includes:

1. Market Researcher
2. Comparative Analyst
3. Requirements Clarifier
4. Recommendation Specialist
5. Industry Expert
6. Technical Evaluator
7. User Experience Specialist
8. Cost-Benefit Analyst
9. Implementation Strategist

Each agent performs specific tasks in the recommendation process, from initial research to final implementation strategies.