Project Title

CrewAi Salesforce Integration

Description

This project integrates Salesforce's AI capabilities with CrewAI, allowing for intelligent routing and handling of user queries.

Salesforce Agent - Build on Coral Cloud Resort Trailhead

Installation

1. Clone the repository.
2. Create a virtual environment:

   python -m venv venv
   

3. Activate the virtual environment:
   • On Windows:

     venv\Scripts\activate
     

   • On macOS/Linux:

     source venv/bin/activate
     

4. Install the required packages using:

   pip install -r requirements.txt
   

Environment Variables

Create a .env file in the project root directory with the following content:

HF_TOKEN=<your_huggingface_token>
SF_API_HOST="https://api.salesforce.com"
SF_ORG_DOMAIN=<your_salesforce_org_domain>
SF_CLIENT_ID=<your_salesforce_client_id>
SF_CLIENT_SECRET=<your_salesforce_client_secret>
SF_AGENT_ID=<your_salesforce_agent_id>

Replace <your_huggingface_token>, <your_salesforce_org_domain>, <your_salesforce_client_id>, <your_salesforce_client_secret>, and <your_salesforce_agent_id> with your actual credentials.

Usage

Run the main application to interact with the Salesforce agent:

python crew_salesforce_tool_app.py

Components

• salesforce_llm_adapter.py: Salesforce Agent Adapter for LLM usage.
• salesforce_crew_llm.py: Wrapper for using Salesforce Agent as LLM.
• salesforce_agent_tool.py: Wrapper to use Salesforce Agent as a tool with CrewAi Agent.
• salesforce_agent_API.py: Handles API interactions with Salesforce.
• crew_salesforce_tool_app.py: Main application for interaction using Salesforce Agent as a tool.
• crew_salesforce_agent_interaction.py: Main application for interactions between Crew AI Agent and Salesforce agent as LLM.

Contributors

• Abhinandan Vijan