Architecture

Role and goal

An agent is a role, a goal, tools, memory, and a model. You do not need every part on day one.

role, goal, and backstory are folded into the system prompt. llm chooses the model.

Example

from praisonai import Agent

agent = Agent(
    name="Researcher",
    role="Senior Research Analyst",
    goal="Uncover developments in AI",
    backstory="You are an expert at a technology research group",
    llm="gpt-4o",
)
print(agent.name)

Try it Yourself

Example explained

  • role, goal, backstory — Who the agent is.
  • llm — Which model it uses, once a key is set.

Reference: documentation