Implementation:CrewAIInc CrewAI Exa Search Tool

Knowledge Sources	CrewAI
Domains	Tools, Web_Search, AI_Services
Last Updated	2026-02-11 00:00 GMT

Overview

Searches the internet using Exa.ai's semantic search API with configurable filters and content retrieval options.

Description

EXASearchTool extends BaseTool to provide semantic internet search through the exa_py library. It initializes an Exa client with optional api_key and base_url (loaded from EXA_API_KEY and EXA_BASE_URL environment variables). The tool supports date range filtering via start_published_date and end_published_date, domain filtering via include_domains, and configurable search type (default: auto). When content mode is enabled, the tool uses search_and_contents() to fetch full page content with optional summary generation. The tool handles automatic installation of the exa_py package via uv if missing.

Usage

Use this tool when a CrewAI agent needs to access real-time internet information through semantic search. Suitable for research agents, fact-checking systems, and knowledge gathering workflows that need current web data with quality controls such as date ranges and domain filtering.

Code Reference

Source Location

Repository: CrewAI
File: lib/crewai-tools/src/crewai_tools/tools/exa_tools/exa_search_tool.py
Lines: 1-135

Signature

class EXASearchTool(BaseTool):
    name: str = "EXASearchTool"
    description: str = "Search the internet using Exa"
    args_schema: type[BaseModel] = EXABaseToolSchema

    client: Any | None = None
    content: bool | None = False
    summary: bool | None = False
    type: str | None = "auto"
    api_key: str | None = ...
    base_url: str | None = ...

    def __init__(self, content: bool | None = False, summary: bool | None = False,
                 type: str | None = "auto", **kwargs): ...
    def _run(self, search_query: str, start_published_date: str | None = None,
             end_published_date: str | None = None,
             include_domains: list[str] | None = None) -> Any: ...

Import

from crewai_tools import EXASearchTool

I/O Contract

Inputs

Name	Type	Required	Description
search_query	str	Yes	The semantic search query
start_published_date	str	No	Start date filter for results
end_published_date	str	No	End date filter for results
include_domains	list[str]	No	List of domains to restrict search results to
content	bool	No	Whether to fetch full page content (constructor param, default: `False`)
summary	bool	No	Whether to include content summaries (constructor param, default: `False`)
type	str	No	Search type (constructor param, default: `auto`)

Outputs

Name	Type	Description
_run() returns	Any	Exa search results object; includes content and summaries if those modes are enabled

Usage Examples

Basic Search

from crewai_tools import EXASearchTool

tool = EXASearchTool()
result = tool.run(search_query="latest advances in AI agents")

Search with Content and Filters

from crewai_tools import EXASearchTool

tool = EXASearchTool(content=True, summary=True)
result = tool.run(
    search_query="machine learning best practices",
    start_published_date="2025-01-01",
    include_domains=["arxiv.org", "openai.com"],
)

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment