Agent Workflow + Research Assistant using AgentQL

In this tutorial, we will use an AgentWorkflow to build a research assistant OpenAI agent using tools including AgentQL's browser tools, Playwright's tools, and the DuckDuckGoSearch tool. This agent performs a web search to find relevant resources for a research topic, interacts with them, and extracts key metadata (e.g., title, author, publication details, and abstract) from those resources.

Initial Setup

The main things we need to get started are:

<a href="https://platform.openai.com/api-keys" target="_blank">OpenAI's API key</a>
<a href="https://dev.agentql.com/api-keys" target="_blank">AgentQL's API key</a>

If you're opening this Notebook on colab, you will probably need to install LlamaIndex 🦙 and Playwright.

python

%pip install llama-index
%pip install llama-index-tools-agentql
%pip install llama-index-tools-playwright
%pip install llama-index-tools-duckduckgo

!playwright install

Store your OPENAI_API_KEY and AGENTQL_API_KEY keys in <a href="https://medium.com/@parthdasawant/how-to-use-secrets-in-google-colab-450c38e3ec75" target="_blank">Google Colab's secrets</a>.

python

import os

from google.colab import userdata

os.environ["AGENTQL_API_KEY"] = userdata.get("AGENTQL_API_KEY")
os.environ["OPENAI_API_KEY"] = userdata.get("OPENAI_API_KEY")

Let's start by enabling async for the notebook since an online environment like Google Colab only supports an asynchronous version of AgentQL.

python

import nest_asyncio

nest_asyncio.apply()

Create an async_browser instance and select the <a href="https://docs.llamaindex.ai/en/latest/api_reference/tools/playwright/" target="_blank">Playwright tools</a> you want to use.

python

from llama_index.tools.playwright.base import PlaywrightToolSpec

async_browser = await PlaywrightToolSpec.create_async_playwright_browser(
    headless=True
)

playwright_tool = PlaywrightToolSpec(async_browser=async_browser)
playwright_tool_list = playwright_tool.to_tool_list()
playwright_agent_tool_list = [
    tool
    for tool in playwright_tool_list
    if tool.metadata.name in ["click", "get_current_page", "navigate_to"]
]

Import the <a href="https://docs.llamaindex.ai/en/latest/api_reference/tools/agentql/" target="_blank">AgentQL browser tools</a> and <a href="https://docs.llamaindex.ai/en/latest/api_reference/tools/duckduckgo/" target="_blank">DuckDuckGo full search tool</a>.

python

from llama_index.tools.agentql import AgentQLBrowserToolSpec
from llama_index.tools.duckduckgo import DuckDuckGoSearchToolSpec

duckduckgo_search_tool = [
    tool
    for tool in DuckDuckGoSearchToolSpec().to_tool_list()
    if tool.metadata.name == "duckduckgo_full_search"
]

agentql_browser_tool = AgentQLBrowserToolSpec(async_browser=async_browser)

We can now create an AgentWorkFlow that uses the tools that we have imported.

python

from llama_index.llms.openai import OpenAI
from llama_index.core.agent.workflow import AgentWorkflow

llm = OpenAI(model="gpt-4o")

workflow = AgentWorkflow.from_tools_or_functions(
    playwright_agent_tool_list
    + agentql_browser_tool.to_tool_list()
    + duckduckgo_search_tool,
    llm=llm,
    system_prompt="You are an expert that can do browser automation, data extraction and text summarization for finding and extracting data from research resources.",
)

AgentWorkflow also supports streaming, which works by using the handler that is returned from the workflow. To stream the LLM output, you can use the AgentStream events.

python

from llama_index.core.agent.workflow import (
    AgentStream,
)

handler = workflow.run(
    user_msg="""
    Use DuckDuckGoSearch to find URL resources on the web that are relevant to the research topic: What is the relationship between exercise and stress levels?
    Go through each resource found. For each different resource, use Playwright to click on link to the resource, then use AgentQL to extract information, including the name of the resource, author name(s), link to the resource, publishing date, journal name, volume number, issue number, and the abstract.
    Find more resources until there are two different resources that can be successfully extracted from.
    """
)

async for event in handler.stream_events():
    if isinstance(event, AgentStream):
        print(event.delta, end="", flush=True)