Praison AI Custom Tools - Integrate ANY Application with AI Agents

---

---

AI Nuggets

Prais and AI Custom Tools Setup Instructions

Installation and Configuration

1. Install Prais and AI:

   pip install prais and Ai  

2. Install DuckDuckGo search package:

   pip install duckduckgo_search  

3. Set environment variables for OpenAI API:
   • Export OpenAI API base URL (Gro URL):

     export OPENAI_API_BASE=<Gro URL>  

   • Export OpenAI model name (Llama 370 billion parameter model):

     export OPENAI_MODEL_NAME=llama  

   • Export OpenAI API key:

     export OPENAI_API_KEY=<Your Gro API Key>  

Creating Agents and Custom Tools

1. Initialize Prais and AI with a task:

   prais and AI -iph init "research about the latest AI news and prepare a detailed report"  

   • This command creates an agents.yaml file with the specified task.
2. Modify the agents.yaml file if needed:
   • Add or edit agents by copying and pasting the agent block and renaming it (e.g., content formatter).
3. Create a Custom Tool:
   • Create a new file named tools.py.
   • Write the following Python code to define an Internet search tool:

     from duckduckgo_search import ddg  
     from prais and AI tools import BaseTool  
      
     class InternetSearchTool(BaseTool):  
         description = "Search internet for relevant information based on a query or latest news"  
      
         def run(self, query):  
             # Integrate your own application or API here  
             return ddg(query)  

   • Add the Custom Tool to the agents.yaml file under the tools section:

     tools:  
       - InternetSearchTool  

Running the Agents

1. Execute Prais and AI:

   prais and AI  

   • This will start the agents and use the Custom Tool to perform tasks.

Using Autogen Framework

1. Run Autogen with Prais and AI:

   prais AI -iph framework autogen  

   • This will run the agents using the Autogen framework.

Integrating with Olo

1. Set the OpenAI model name for Olo:

   export OPENAI_MODEL_NAME=llama3  

2. Set the OpenAI API base for Olo:

   export OPENAI_API_BASE=http://localhost:11434/slv1  

3. Pull the Llama 3 Model using AMA:

   ama pull llama3  

4. Run Prais and AI with Olo integration:

   prais and AI  

Additional Tips

• Make sure to replace <Gro URL> and <Your Gro API Key> with your actual Gro URL and API key.
• The Llama 3 model is smaller and may not be as effective as larger models for certain tasks.
• You might need a larger model or a fine-tuned version of Llama 3 for better performance.

Video and Channel Information

• The instructions are based on a YouTube video from the channel that regularly creates content about Artificial Intelligence.
• Viewers are encouraged to subscribe, click the bell icon, like, share, and subscribe for more similar content.

For more detailed steps and explanations, refer to the video at http://youtube.com/watch?v=JSU2Rndh06c.