ExaSearchTool - Unable to get it configured

CrewAI Community Support
,
1

Hi, can anyone help me with a simple code snippet to get CrewAI’s ExaSearchTool to work. I am really struggling use the BaseModel and the decorator approach and can’t figure it out.

2

Btw. I have tried this code snippet but the Crew agent keeps gettng the structure of the call to Exa_search wrong: from crewai_tools import tool
from exa_py import Exa
import os
from typing import Union, Dict, Any

@tool
def exa_search(query: str) → str:
“”"
Search the web using EXA API with optimized search parameters.
“”"
try:
# Execute search
api_key = os.getenv(“EXA_API_KEY”)
if not api_key:
return “Error: EXA API key not found in environment variables”

    exa = Exa(api_key=api_key)
    response = exa.search_and_contents(
        query=query,
        num_results=5,
        use_autoprompt=True,
        type="auto",
        text=True,
        highlights=True,
        summary=True
    )

    # Format results
    formatted_results = []
    for result in response.results:
        sections = [
            f"Title: {result.title or 'No title available'}",
            f"Link: {result.url}",
            "Snippet:"
        ]

        if hasattr(result, 'summary') and result.summary:
            sections.append("\nSummary:")
            sections.append(f"  {result.summary}")

        if hasattr(result, 'highlights') and result.highlights:
            sections.append("\nKey Points:")
            for highlight in result.highlights:
                sections.append(f"  • {highlight}")

        formatted_results.append("\n".join(sections) + "\n\n---\n")

    return "\n".join(formatted_results)

except Exception as e:
    return f"Error: {str(e)}"
3

exa didnt work for me too it doesnt even recognize the exa_py in the import

4

Try closing and reopening your IDE. This often resolves errors with unrecognized installed packages.

5

Try this way

pip install exa_py crewai

import os
from crewai import Agent, Task, Crew
from exa_py import Exa

Initialize Exa client

exa_client = Exa(api_key=os.environ.get(“EXA_API_KEY”))

Create a researcher agent that uses Exa

researcher = Agent(
role=“Research Analyst”,
goal=“Find accurate and up-to-date information”,
backstory=“Expert at gathering information from the web”,
tools=[
{
“name”: “exa_search”,
“description”: “Search the web for information”,
“func”: lambda query: exa_search(query, exa_client)
}
]
)

Define the search function

def exa_search(query, client):
results = client.search_and_contents(
query=query,
num_results=5,
text={“max_characters”: 3000}
)

formatted_results = []
for result in results.results:
    formatted_results.append({
        "title": result.title,
        "url": result.url,
        "text": result.text[:3000] if result.text else "No text available"
    })

return formatted_results

Create tasks and crew as usual in CrewAI

task = Task(
description=“Research recent developments in quantum computing”,
agent=researcher
)

crew = Crew(
agents=[researcher],
tasks=[task]
)

result = crew.kickoff()