Skip to content

Python SDK for AI agent monitoring, LLM cost tracking, benchmarking, and more. Integrates with most LLMs and agent frameworks including CrewAI, Agno, OpenAI Agents SDK, Langchain, Autogen, AG2, and CamelAI

5.3 KB
Raw Blame
---
title: 'Smolagents'
description: 'Orchestrate a Multi-Agent System'
---
{/*  SOURCE_FILE: examples/smolagents/multi_smolagents_system.ipynb  */}

_View Notebook on <a href={'https://github.com/AgentOps-AI/agentops/blob/main/examples/smolagents/multi_smolagents_system.ipynb'} target={'_blank'}>Github</a>_

# Orchestrate a Multi-Agent System

In this notebook, we will make a multi-agent web browser: an agentic system with several agents collaborating to solve problems using the web!

It will be a simple hierarchy, using a `ManagedAgent` object to wrap the managed web search agent:

```
+----------------+
| Manager agent  |
+----------------+
         |
_________|______________
|                        |
Code interpreter   +--------------------------------+
       tool        |         Managed agent          |
                   |      +------------------+      |
                   |      | Web Search agent |      |
                   |      +------------------+      |
                   |         |            |         |
                   |  Web Search tool     |         |
                   |             Visit webpage tool |
                   +--------------------------------+
```
Let’s set up this system.

Run the line below to install the required dependencies:



## Installation
<CodeGroup>
  ```bash pip
  pip install agentops duckduckgo-search markdownify smolagents
  ```
  ```bash poetry
  poetry add agentops duckduckgo-search markdownify smolagents
  ```
  ```bash uv
  uv pip install agentops duckduckgo-search markdownify smolagents
  ```
</CodeGroup>

🖇️ Now we initialize the AgentOps client and load the environment variables to use the API keys.


```python
import agentops
from dotenv import load_dotenv
import os
import re
import requests
from markdownify import markdownify
from requests.exceptions import RequestException
from smolagents import LiteLLMModel, tool, CodeAgent, ToolCallingAgent, DuckDuckGoSearchTool

load_dotenv()
os.environ["AGENTOPS_API_KEY"] = os.getenv("AGENTOPS_API_KEY", "your_api_key_here")
os.environ["OPENAI_API_KEY"] = os.getenv("OPENAI_API_KEY", "your_openai_api_key_here")
```

⚡️ Our agent will be powered by `openai/gpt-4o-mini` using the `LiteLLMModel` class.


```python
from smolagents import LiteLLMModel, tool ,CodeAgent, ToolCallingAgent, DuckDuckGoSearchTool
agentops.init(auto_start_session=False)
tracer = agentops.start_trace(
    trace_name="Orchestrate a Multi-Agent System", tags=["smolagents", "example", "multi-agent", "agentops-example"]
)
model = LiteLLMModel("openai/gpt-4o-mini")
```

## Create a Web Search Tool

For web browsing, we can already use our pre-existing `DuckDuckGoSearchTool`. However, we will also create a `VisitWebpageTool` from scratch using `markdownify`. Here’s how:


```python
@tool
def visit_webpage(url: str) -> str:
    """Visits a webpage at the given URL and returns its content as a markdown string.

    Args:
        url: The URL of the webpage to visit.

    Returns:
        The content of the webpage converted to Markdown, or an error message if the request fails.
    """
    try:
        # Send a GET request to the URL
        response = requests.get(url)
        response.raise_for_status()  # Raise an exception for bad status codes

        # Convert the HTML content to Markdown
        markdown_content = markdownify(response.text).strip()

        # Remove multiple line breaks
        markdown_content = re.sub(r"\n{3,}", "\n\n", markdown_content)

        return markdown_content

    except RequestException as e:
        return f"Error fetching the webpage: {str(e)}"
    except Exception as e:
        return f"An unexpected error occurred: {str(e)}"
```

Let’s test our tool:


```python
print(visit_webpage("https://en.wikipedia.org/wiki/Hugging_Face")[:500])
```

## Build Our Multi-Agent System

We will now use the tools `search` and `visit_webpage` to create the web agent.


```python
web_agent = ToolCallingAgent(
    tools=[DuckDuckGoSearchTool(), visit_webpage],
    model=model,
    name="search",
    description="Runs web searches for you. Give it your query as an argument.",
)

manager_agent = CodeAgent(
    tools=[],
    model=model,
    managed_agents=[web_agent],
    additional_authorized_imports=["time", "numpy", "pandas"],
)
```

Let’s run our system with the following query:


```python
answer = manager_agent.run(
    "If LLM trainings continue to scale up at the current rhythm until 2030, what would be the electric power in GW required to power the biggest training runs by 2030? What does that correspond to, compared to some countries? Please provide a source for any number used."
)

print(answer)
```

Awesome! We've successfully run a multi-agent system. Let's end the agentops session with a "Success" state. You can also end the session with a "Failure" or "Indeterminate" state, which is set as default.


```python
agentops.end_trace(tracer, end_state="Success")
```

You can view the session in the [AgentOps dashboard](https://app.agentops.ai/sessions) by clicking the link provided after ending the session.


<script type="module" src="/scripts/github_stars.js"></script>
<script type="module" src="/scripts/scroll-img-fadein-animation.js"></script>
<script type="module" src="/scripts/button_heartbeat_animation.js"></script>
<script type="module" src="/scripts/adjust_api_dynamically.js"></script>