Implementation:CrewAIInc CrewAI Serper Dev Tool

Knowledge Sources	CrewAI
Domains	Web Search, Tool Integration, Information Retrieval
Last Updated	2026-02-11 00:00 GMT

Overview

SerperDevTool is a CrewAI tool that performs internet searches via the Serper.dev API, supporting both general web search and news search with structured, typed result processing.

Description

The tool extends BaseTool and provides configurable internet search functionality through the Serper.dev Google Search API. It defines multiple TypedDict classes for strongly typed response data:

• KnowledgeGraph -- Title, type, website, image, description, and attributes from Google's knowledge panel.
• OrganicResult -- Title, link, snippet, position, and sitelinks for standard search results.
• PeopleAlsoAskResult -- Question, snippet, title, and link for "People Also Ask" results.
• RelatedSearchResult -- Related search queries.
• NewsResult -- Title, link, snippet, date, source, and image for news results.
• FormattedResults -- The top-level container combining all result types.

The _run method orchestrates the search workflow: it validates the search query, calls _make_api_request to POST to the appropriate Serper endpoint (https://google.serper.dev/search or /news) with the API key from SERPER_API_KEY, then delegates to type-specific processors that extract and normalize fields, limiting results to n_results. Configurable country, location, and locale filters allow for geo-targeted searches. Results can optionally be saved to timestamped files.

Usage

Use this tool when agents need to perform internet searches, gather web intelligence, retrieve news articles, or find information about topics, companies, products, or people. It supports two search types: "search" for general web results and "news" for news-specific results.

Code Reference

Source Location

• Repository: CrewAI
• File: lib/crewai-tools/src/crewai_tools/tools/serper_dev_tool/serper_dev_tool.py
• Lines: 1-343

Signature

class SerperDevTool(BaseTool):
    name: str = "Search the internet with Serper"
    description: str = (
        "A tool that can be used to search the internet with a search_query. "
        "Supports different search types: 'search' (default), 'news'"
    )
    args_schema: type[BaseModel] = SerperDevToolSchema
    base_url: str = "https://google.serper.dev"
    n_results: int = 10
    save_file: bool = False
    search_type: str = "search"
    country: str | None = ""
    location: str | None = ""
    locale: str | None = ""
    env_vars: list[EnvVar]  # Requires SERPER_API_KEY

Import

from crewai_tools.tools.serper_dev_tool.serper_dev_tool import SerperDevTool

I/O Contract

Inputs

Name	Type	Required	Description
search_query	str	Yes	The search query to use for internet search
search_type	str	No	Type of search: "search" (default) or "news"
save_file	bool	No	Whether to save results to a timestamped file (default: False)

Constructor Parameters

Name	Type	Required	Description
base_url	str	No	Base URL for Serper API (default: "https://google.serper.dev")
n_results	int	No	Maximum number of results to return (default: 10)
save_file	bool	No	Whether to save results to file by default (default: False)
search_type	str	No	Default search type: "search" or "news" (default: "search")
country	str	No	Country code for geo-targeted results (e.g., "us")
location	str	No	Location string for geo-targeted results
locale	str	No	Locale code for language-specific results (e.g., "en")

Outputs

Name	Type	Description
return	FormattedResults	Dictionary containing searchParameters, organic results, knowledgeGraph, peopleAlsoAsk, relatedSearches, news (for news search type), and credits used

Usage Examples

Basic Usage

import os
os.environ["SERPER_API_KEY"] = "your-serper-api-key"

from crewai_tools.tools.serper_dev_tool.serper_dev_tool import SerperDevTool

# General web search
tool = SerperDevTool(n_results=5)
results = tool._run(search_query="CrewAI multi-agent framework")

# News search with geo-targeting
tool = SerperDevTool(search_type="news", country="us", n_results=10)
results = tool._run(search_query="artificial intelligence breakthroughs")

# Save results to file
tool = SerperDevTool(save_file=True)
results = tool._run(search_query="Python best practices 2025")