When building an Agent, it is important to consider the following:
  • Which model to use? -> See the Models documentation.
  • Which tools to use? -> See the Tools documentation.
  • Which instructions to use? -> See the Context Engineering documentation.
  • Should I enable reasoning? -> See the Reasoning documentation.
  • Do I need to add knowledge? -> See the Knowledge documentation.
  • How do I configure storage? -> See the Storage documentation.
  • How do I configure memory? -> See the Memory documentation.

​
The Development Process

It is typical while developing an Agent to use the Agent.print_response() method to print the response in the terminal. This is only for convenience during development and not recommended for production use. 
from agno.agent import Agent
from agno.models.openai import OpenAIChat

agent = Agent(model=OpenAIChat(id="gpt-5-mini"))

agent.print_response("Tell me a 5 second short story about a robot")

# Or for streaming
agent.print_response("Tell me a 5 second short story about a robot", stream=True)

# Or async
await agent.aprint_response("Tell me a 5 second short story about a robot")

# Or for async streaming
await agent.aprint_response("Tell me a 5 second short story about a robot", stream=True)
The Agent.print_response() and Agent.aprint_response() methods are helper methods that use the Agent.run() and Agent.arun() methods under the hood.See Running your Agent for more details.
See the full method signature in the Agent class reference.

​
Debugging

You can enable debug mode to get more detailed logs. 
agent = Agent(model=OpenAIChat(id="gpt-5-mini"), debug_mode=True)

# You can also set it directly on the run
agent.print_response("Tell me a 5 second short story about a robot", debug_mode=True)

​
Interactive CLI

You can also interact with the agent via a CLI. 
agent.cli_app(input="Tell me a 5 second short story about a robot", stream=True)
See the Agent class reference for more details.

​
Developer Resources