View All Notebooks on Github

​
MultiOn Tracking Web Browse Actions

Agents using MultiOn can launch and control remote or local web browsers to perform actions and retrieve context using natural language commands. With AgentOps, MultiOn evens such as browse, retrieve, and step are automatically tracked.

Furthermore, events and LLM calls in your Python program will be tracked as well.

First let’s install the required packages

%pip install -U multion
%pip install -U agentops
%pip install -U openai
%pip install -U python-dotenv

Then import them

from multion.client import MultiOn
from multion.core.request_options import RequestOptions
import openai
import agentops
import os
from dotenv import load_dotenv

Next, we’ll set our API keys. There are several ways to do this, the code below is just the most foolproof way for the purposes of this notebook. It accounts for both users who use environment variables and those who just want to set the API Key here in this notebook.

Get an AgentOps API key

  1. Create an environment variable in a .env file or other method. By default, the AgentOps init() function will look for an environment variable named AGENTOPS_API_KEY. Or…

  2. Replace <your_agentops_key> below and pass in the optional api_key parameter to the AgentOps init(api_key=...) function. Remember not to commit your API key to a public repo!

load_dotenv()
MULTION_API_KEY = os.getenv("MULTION_API_KEY") or "<your_multion_key>"
AGENTOPS_API_KEY = os.getenv("AGENTOPS_API_KEY") or "<your_agentops_key>"
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY") or "<your_openai_key>"

​
Tracking MultiOn events with AgentOps

When an agentops_api_key is provided, MultiOn will automatically start an AgentOps session and record events.

multion = MultiOn(
    api_key=MULTION_API_KEY,
    agentops_api_key=AGENTOPS_API_KEY,
)
cmd = "what three things do i get with agentops"
request_options = RequestOptions(
    timeout_in_seconds=60, max_retries=4, additional_headers={"test": "ing"}
)

browse_response = multion.browse(
    cmd="what three things do i get with agentops",
    url="https://www.agentops.ai/",
    max_steps=4,
    include_screenshot=True,
    request_options=request_options,
)

print(browse_response.message)

# End session to see your dashboard
agentops.end_session("Success")

​
Linking MultiOn events to an existing AgentOps session

When running agentops.init(), be sure to set auto_start_session=False. MultiOn will automatically launch AgentOps sessions by default, but by setting auto start to false, you can configure your AgentOps client independently.

agentops.init(
    AGENTOPS_API_KEY, auto_start_session=False, default_tags=["MultiOn browse example"]
)

Now, we can launch a MultiOn browse event. This event will automatically get added to your AgentOps session.

multion = MultiOn(
    api_key=MULTION_API_KEY,
    agentops_api_key=AGENTOPS_API_KEY,
)
cmd = "what three things do i get with agentops"
request_options = RequestOptions(
    timeout_in_seconds=60, max_retries=4, additional_headers={"test": "ing"}
)

browse_response = multion.browse(
    cmd="what three things do i get with agentops",
    url="https://www.agentops.ai/",
    max_steps=4,
    include_screenshot=True,
    request_options=request_options,
)

print(browse_response.message)

Let’s use OpenAI to summarize our output

messages = [
    {
        "role": "user",
        "content": f"Format this data as a markdown table: {browse_response.message}",
    }
]
client = openai.OpenAI()
response = client.chat.completions.create(messages=messages, model="gpt-3.5-turbo")

print(response.choices[0].message.content)

agentops.end_session("Success")

​
Check your session

Check your session on AgentOps. This session should include the MultiOn browse action and the OpenAI call.