Skip to main content
Prerequisites: Basic understanding of Agno and completed Deploy Your First Agent tutorial

​
Overview

Agno is a framework for building AI agents with print response capabilities and streaming support. RunAgent makes it easy to deploy Agno agents and access them from any programming language while maintaining the framework’s unique characteristics.

​
Quick Start

​
1. Create an Agno Agent

runagent init my-agno-agent --framework agno
cd my-agno-agent

​
2. Install Dependencies

pip install -r requirements.txt

​
3. Configure Your Agent

The generated runagent.config.json will be pre-configured for Agno: 
{
  "agent_name": "my-agno-agent",
  "description": "Agno AI agent with print capabilities",
  "framework": "agno",
  "agent_architecture": {
    "entrypoints": [
      {
        "file": "agent.py",
        "module": "agno_agent",
        "tag": "main"
      }
    ]
  }
}

​
Basic Agno Agent

Here’s a simple Agno agent that demonstrates the core concepts:
agent.py
from agno import Agent, Tool
from typing import Dict, Any, List, Optional
import json
from datetime import datetime

# Define tools
def calculate(expression: str) -> str:
    """Calculate mathematical expressions safely"""
    try:
        # Simple safe evaluation (in production, use a proper math parser)
        allowed_chars = set('0123456789+-*/.() ')
        if all(c in allowed_chars for c in expression):
            result = eval(expression)
            return f"Result: {result}"
        else:
            return "Error: Invalid characters in expression"
    except Exception as e:
        return f"Error: {str(e)}"

def get_current_time() -> str:
    """Get current time"""
    return datetime.now().strftime("%Y-%m-%d %H:%M:%S")

def format_text(text: str, format_type: str = "uppercase") -> str:
    """Format text in different ways"""
    if format_type == "uppercase":
        return text.upper()
    elif format_type == "lowercase":
        return text.lower()
    elif format_type == "title":
        return text.title()
    else:
        return text

# Create tools
tools = [
    Tool(name="calculate", function=calculate, description="Calculate mathematical expressions"),
    Tool(name="get_time", function=get_current_time, description="Get current time"),
    Tool(name="format_text", function=format_text, description="Format text in different ways")
]

# Create Agno agent
agno_agent = Agent(
    name="AgnoAssistant",
    tools=tools,
    system_message="You are a helpful Agno assistant. You can perform calculations, get time, and format text. Always show your work step by step."
)

def agno_agent(message: str, user_id: str = "default") -> Dict[str, Any]:
    """Main entrypoint for the Agno agent"""
    try:
        # Process message with Agno
        response = agno_agent.process_message(message)
        
        return {
            "response": response,
            "user_id": user_id,
            "timestamp": datetime.now().isoformat(),
            "status": "success"
        }
        
    except Exception as e:
        return {
            "response": f"Error: {str(e)}",
            "user_id": user_id,
            "timestamp": datetime.now().isoformat(),
            "status": "error"
        }

​
Advanced Agno Patterns

​
1. Streaming Agno Agent

streaming_agno.py
from typing import Iterator, Dict, Any
from agno import Agent, Tool

def streaming_agno_agent(message: str, user_id: str = "default") -> Iterator[str]:
    """Streaming Agno agent with step-by-step output"""
    yield f"🤖 Agno Agent starting processing...\n\n"
    
    # Initialize agent
    agent = Agent(
        name="StreamingAgnoAssistant",
        tools=tools,
        system_message="You are a helpful Agno assistant that shows your work step by step."
    )
    
    yield f"📝 Processing message: {message}\n\n"
    
    # Simulate step-by-step processing
    yield "🔍 Analyzing request...\n"
    yield "🛠️ Selecting appropriate tools...\n"
    yield "⚙️ Executing operations...\n"
    yield "📊 Formatting results...\n\n"
    
    # Process message
    response = agent.process_message(message)
    
    yield f"✅ Response: {response}\n\n"
    yield "🎉 Processing complete!"

​
2. Specialized Agno Agents

specialized_agno.py
from agno import Agent, Tool
from typing import Dict, Any

# Math Agent
def math_agent(problem: str, user_id: str) -> Dict[str, Any]:
    """Math-focused Agno agent"""
    math_tools = [
        Tool(name="calculate", function=calculate, description="Calculate mathematical expressions"),
        Tool(name="solve_equation", function=lambda x: f"Solving equation: {x}", description="Solve equations"),
        Tool(name="convert_units", function=lambda x: f"Converting units: {x}", description="Convert between units")
    ]
    
    agent = Agent(
        name="MathAssistant",
        tools=math_tools,
        system_message="You are a math assistant. Show all steps and explain your reasoning."
    )
    
    response = agent.process_message(problem)
    
    return {
        "response": response,
        "agent_type": "math",
        "user_id": user_id,
        "status": "success"
    }

# Text Processing Agent
def text_agent(text: str, operation: str, user_id: str) -> Dict[str, Any]:
    """Text processing Agno agent"""
    text_tools = [
        Tool(name="format_text", function=format_text, description="Format text in different ways"),
        Tool(name="count_words", function=lambda x: f"Word count: {len(x.split())}", description="Count words"),
        Tool(name="extract_keywords", function=lambda x: f"Keywords: {x.split()[:5]}", description="Extract keywords")
    ]
    
    agent = Agent(
        name="TextAssistant",
        tools=text_tools,
        system_message="You are a text processing assistant. Help with text manipulation and analysis."
    )
    
    response = agent.process_message(f"Process this text: {text} with operation: {operation}")
    
    return {
        "response": response,
        "agent_type": "text",
        "user_id": user_id,
        "status": "success"
    }

# Data Analysis Agent
def data_agent(data: str, analysis_type: str, user_id: str) -> Dict[str, Any]:
    """Data analysis Agno agent"""
    data_tools = [
        Tool(name="calculate", function=calculate, description="Calculate mathematical expressions"),
        Tool(name="statistics", function=lambda x: f"Statistics for: {x}", description="Calculate statistics"),
        Tool(name="visualize", function=lambda x: f"Visualization: {x}", description="Create data visualizations")
    ]
    
    agent = Agent(
        name="DataAssistant",
        tools=data_tools,
        system_message="You are a data analysis assistant. Help analyze and visualize data."
    )
    
    response = agent.process_message(f"Analyze this data: {data} with analysis type: {analysis_type}")
    
    return {
        "response": response,
        "agent_type": "data",
        "user_id": user_id,
        "status": "success"
    }

​
3. Multi-Step Agno Workflow

workflow_agno.py
from agno import Agent, Tool
from typing import Dict, Any, List

def multi_step_agno_agent(task: str, user_id: str) -> Dict[str, Any]:
    """Multi-step Agno workflow agent"""
    
    # Step 1: Planning Agent
    planning_tools = [
        Tool(name="break_down_task", function=lambda x: f"Breaking down: {x}", description="Break down complex tasks"),
        Tool(name="identify_requirements", function=lambda x: f"Requirements: {x}", description="Identify requirements")
    ]
    
    planning_agent = Agent(
        name="PlanningAgent",
        tools=planning_tools,
        system_message="You are a planning agent. Break down complex tasks into manageable steps."
    )
    
    # Step 2: Execution Agent
    execution_tools = [
        Tool(name="calculate", function=calculate, description="Calculate mathematical expressions"),
        Tool(name="format_text", function=format_text, description="Format text in different ways"),
        Tool(name="get_time", function=get_current_time, description="Get current time")
    ]
    
    execution_agent = Agent(
        name="ExecutionAgent",
        tools=execution_tools,
        system_message="You are an execution agent. Execute the planned steps."
    )
    
    # Step 3: Review Agent
    review_tools = [
        Tool(name="check_accuracy", function=lambda x: f"Checking accuracy: {x}", description="Check accuracy"),
        Tool(name="format_output", function=lambda x: f"Formatting output: {x}", description="Format final output")
    ]
    
    review_agent = Agent(
        name="ReviewAgent",
        tools=review_tools,
        system_message="You are a review agent. Review and improve the results."
    )
    
    # Execute workflow
    try:
        # Planning phase
        plan = planning_agent.process_message(f"Plan this task: {task}")
        
        # Execution phase
        execution = execution_agent.process_message(f"Execute this plan: {plan}")
        
        # Review phase
        review = review_agent.process_message(f"Review this result: {execution}")
        
        return {
            "task": task,
            "plan": plan,
            "execution": execution,
            "review": review,
            "user_id": user_id,
            "status": "success"
        }
        
    except Exception as e:
        return {
            "task": task,
            "error": str(e),
            "user_id": user_id,
            "status": "error"
        }

​
Configuration for Multiple Agents

Update your runagent.config.json to include multiple Agno agents: 
{
  "agent_name": "advanced-agno-agent",
  "description": "Advanced Agno multi-agent system",
  "framework": "agno",
  "agent_architecture": {
    "entrypoints": [
      {
        "file": "agent.py",
        "module": "agno_agent",
        "tag": "main"
      },
      {
        "file": "streaming_agno.py",
        "module": "streaming_agno_agent",
        "tag": "streaming"
      },
      {
        "file": "specialized_agno.py",
        "module": "math_agent",
        "tag": "math"
      },
      {
        "file": "specialized_agno.py",
        "module": "text_agent",
        "tag": "text"
      },
      {
        "file": "specialized_agno.py",
        "module": "data_agent",
        "tag": "data"
      },
      {
        "file": "workflow_agno.py",
        "module": "multi_step_agno_agent",
        "tag": "workflow"
      }
    ]
  }
}

​
Testing Your Agno Agent

​
Python Client

test_agno.py
from runagent import RunAgentClient

# Connect to your Agno agent
client = RunAgentClient(
    agent_id="your_agent_id_here",
    entrypoint_tag="main",
    local=True
)

# Test basic functionality
result = client.run(message="Calculate 2 + 2 * 3", user_id="test_user")
print(f"Math Result: {result['response']}")

# Test text processing
text_client = RunAgentClient(
    agent_id="your_agent_id_here",
    entrypoint_tag="text",
    local=True
)

text_result = text_client.run(
    text="hello world",
    operation="uppercase",
    user_id="test_user"
)
print(f"Text Result: {text_result['response']}")

# Test streaming
stream_client = RunAgentClient(
    agent_id="your_agent_id_here",
    entrypoint_tag="streaming",
    local=True
)

print("Streaming Agno workflow:")
for chunk in stream_client.run(message="Calculate the area of a circle with radius 5", user_id="test_user"):
    print(chunk, end="", flush=True)

​
JavaScript Client

test_agno.js
import { RunAgentClient } from 'runagent';

const client = new RunAgentClient({
    agentId: 'your_agent_id_here',
    entrypointTag: 'main',
    local: true
});

await client.initialize();

const result = await client.run({
    message: 'Calculate 2 + 2 * 3',
    user_id: 'test_user'
});

console.log('Math Result:', result.response);

​
Best Practices

​
1. Tool Design

  • Create focused, single-purpose tools
  • Provide clear tool descriptions
  • Handle tool errors gracefully

​
2. Agent Configuration

  • Use appropriate system messages
  • Configure agent behavior clearly
  • Test agent responses

​
3. Error Handling

  • Implement try-catch blocks
  • Provide meaningful error messages
  • Log errors for debugging

​
4. Performance

  • Optimize tool execution
  • Use caching when appropriate
  • Monitor agent performance

​
Common Patterns

Use Agno agents for mathematical calculations and problem-solving.
Create agents specialized in text processing and formatting.
Build agents for data analysis and visualization.
Implement complex workflows with multiple agent steps.

​
Troubleshooting

​
Common Issues

  1. Tool Execution Errors
    • Check tool definitions
    • Verify tool permissions
    • Handle tool errors
  2. Agent Configuration
    • Verify system messages
    • Check agent settings
    • Test agent responses
  3. Performance Issues
    • Monitor tool execution time
    • Optimize agent configuration
    • Use caching when appropriate

​
Debug Tips

# Add debugging to your Agno agents
def debug_agno_agent(message: str, user_id: str) -> Dict[str, Any]:
    print(f"Debug: Processing message from {user_id}: {message}")
    
    # Your agent logic here
    
    print(f"Debug: Response generated: {response}")
    return result

​
Performance Optimization

​
1. Tool Optimization

  • Cache expensive tool results
  • Optimize tool execution
  • Use async tools when possible

​
2. Agent Configuration

  • Optimize system messages
  • Configure appropriate timeouts
  • Use efficient agent settings

​
3. Workflow Optimization

  • Minimize unnecessary steps
  • Use parallel processing when possible
  • Optimize data flow

​
Next Steps

Advanced Patterns

Learn advanced Agno patterns and techniques

Production Deployment

Deploy your Agno agents to production

Multi-Language Access

Access your Agno agents from different languages

Performance Tuning

Optimize your Agno agents for production
🎉 Excellent work! You’ve learned how to deploy Agno agents with RunAgent. Agno’s unique print response capabilities combined with RunAgent’s multi-language access create powerful, transparent AI systems!