Navigate to Agent Builder:
From the Lyzr Agent Studio page, select Build Agent.
Provide Agent Details:
Enter a unique name for your agent and add a detailed description outlining its purpose.
Select LLM Model:
Choose an appropriate Large Language Model provider and specific model for your agent’s use case. Configure parameters like temperature and Top P as needed.
Define Agent Role and Instructions:
Specify the agent’s role (e.g., Customer Support, Data Fetcher) and provide instructions to guide its behavior. Use the Improve feature to refine these details.
Add Tools (Optional):
Enhance your agent’s capabilities by integrating tools such as email sending, database queries, or custom APIs.
Add Examples and Knowledge Bases (Optional):
Provide example prompts and responses to train your agent, and link relevant knowledge bases to improve accuracy.
Select Features:
Choose from available features to further customize the agent’s behavior and output.
Create and Test Agent:
Click Create Agent to save. Test the agent with sample queries and fine-tune as needed.
Deploy Agent (Optional):
Deploy your agent by publishing it to the Agent Marketplace or use the provided JSON and API key for integration into your applications.