Skip to main content

Lab 12: Building a Research Assistant with Web Search Challenge

Goal​

Goal​

In this lab, you will build a Research Assistant that can access up-to-date information from the internet. You will learn how to use the GoogleSearchAgentTool wrapper to combine web search with your own custom tools.

Prerequisites​

  • Note on Costs: This lab requires the Vertex AI API, which may incur small costs on your Google Cloud bill.

Step 1: Create and Prepare the Project​

  1. Create the agent project:

    adk create research-assistant

    When prompted, choose the Programmatic (Python script) option.

  2. Navigate into the new directory:

    cd research-assistant

  3. Create the .env file: The GoogleSearchAgentTool requires a Vertex AI configuration. Create a .env file in this directory with the following content, replacing <your_gcp_project> with your actual Google Cloud project ID.

    GOOGLE_GENAI_USE_VERTEXAI=1
    GOOGLE_CLOUD_PROJECT=<your_gcp_project>
    GOOGLE_CLOUD_LOCATION=us-central1

Step 2: Define the Agent and Tools​

Exercise: Open agent.py. The custom tools have been provided for you. Your task is to complete the agent definition by instantiating the search tool and writing the agent's instructions.

# In agent.py (Starter Code)

from datetime import datetime
from google.adk.agents import Agent
from google.adk.tools import FunctionTool, GoogleSearchAgentTool

# --- Custom Tools (Provided for you) ---

def format_research_notes(topic: str, findings: str) -> dict:
    """Formats research findings into a document."""
    timestamp = datetime.now().strftime('%Y-%m-%d %H:%M:%S')
    document = f"""
# Research Report: {topic}
Generated: {timestamp}

## Findings
{findings}
    """.strip()
    return {"status": "success", "document": document}

def extract_key_facts(text: str, num_facts: int = 5) -> list[str]:
    """Extract key facts from text (simplified)."""
    sentences = text.split('.')
    return [s.strip() for s in sentences if s.strip()][:num_facts]

# --- Agent Definition ---

# TODO: 1. Create an instance of the GoogleSearchAgentTool.
search_tool = ...

# TODO: 2. Define the `root_agent`.
root_agent = Agent(
    model='gemini-2.5-flash',
    name='research_assistant',
    description='Conducts web research and compiles findings',
    instruction="""
    # TODO: 3. Write an instruction that tells the agent to perform the research
    # workflow in the correct order:
    # 1. Use search_tool to find information.
    # 2. Use extract_key_facts on the search results.
    # 3. Use format_research_notes on the extracted facts.
    # 4. Present the final document as the answer.
    """,
    tools=[
        # TODO: 4. Add the `search_tool` and the two custom tools
        # (`extract_key_facts`, `format_research_notes`) to this list.
    ]
)

Step 3: Run and Test the Research Assistant​

  1. Navigate to the parent directory:

    cd ..

  2. Start the Dev UI:

    adk web research-assistant

  3. Interact with the agent:

    • Give the agent a research topic, like "What are the latest developments in AI in 2025?"

  4. Analyze the Trace View:

    • Expand the trace for your query. You should see a sequence of tool calls: GoogleSearchAgentTool, then extract_key_facts, then format_research_notes.

Having Trouble?​

If you get stuck, you can find the complete, working code in the lab-solution.md file.

Lab Summary​

You have successfully built an agent that can access real-world, current information from the internet and process it using custom logic. You have learned:

  • How to use the GoogleSearchAgentTool wrapper to combine built-in search with custom FunctionTools.
  • How to write an instruction that orchestrates a sequence of different tool types.

Self-Reflection Questions​

  • The GoogleSearchAgentTool is a workaround for a current limitation. Why is it architecturally cleaner to have a wrapper like this instead of building the search logic directly into your own custom tool?
  • Our extract_key_facts tool is very simple. How could you make it more robust? (Hint: Could another LLM be used for this task?)
  • The agent's instruction defines a specific, sequential workflow. What might happen if you didn't specify the order of the tool calls in the instruction?

🕵️ Hidden Solution 🕵️​

Looking for the solution? Here's a hint (Base64 decode me): L2RvYy1hZGstdHJhaW5pbmcvbW9kdWxlMTItYnVpbHQtaW4tdG9vbHMtZ3JvdW5kaW5nL2xhYi1zb2x1dGlvbg==

The direct link is: Lab Solution