BrightDataSERP
Bright Data provides a powerful SERP API that allows you to query search engines (Google,Bing.DuckDuckGo,Yandex) with geo-targeting and advanced customization options, particularly useful for AI agents requiring real-time web information.
Overviewโ
Integration detailsโ
Class | Package | Serializable | JS support | Package latest |
---|---|---|---|---|
BrightDataSERP | langchain-brightdata | โ | โ |
Tool featuresโ
Native async | Returns artifact | Return data | Pricing |
---|---|---|---|
โ | โ | Title, URL, snippet, position, and other search result data | Requires Bright Data account |
Setupโ
The integration lives in the langchain-brightdata
package.
pip install langchain-brightdata
Credentialsโ
You'll need a Bright Data API key to use this tool. You can set it as an environment variable:
import os
os.environ["BRIGHT_DATA_API_KEY"] = "your-api-key"
Or pass it directly when initializing the tool:
from langchain_brightdata import BrightDataSERP
serp_tool = BrightDataSERP(bright_data_api_key="your-api-key")
Instantiationโ
Here we show how to instantiate an instance of the BrightDataSERP tool. This tool allows you to perform search engine queries with various customization options including geo-targeting, language preferences, device type simulation, and specific search types using Bright Data's SERP API.
The tool accepts various parameters during instantiation:
bright_data_api_key
(required, str): Your Bright Data API key for authentication.search_engine
(optional, str): Search engine to use for queries. Default is "google". Other options include "bing", "yahoo", "yandex", "DuckDuckGo" etc.country
(optional, str): Two-letter country code for localized search results (e.g., "us", "gb", "de", "jp"). Default is "us".language
(optional, str): Two-letter language code for the search results (e.g., "en", "es", "fr", "de"). Default is "en".results_count
(optional, int): Number of search results to return. Default is 10. Maximum value is typically 100.search_type
(optional, str): Type of search to perform. Options include:- None (default): Regular web search
- "isch": Images search
- "shop": Shopping search
- "nws": News search
- "jobs": Jobs search
device_type
(optional, str): Device type to simulate for the search. Options include:- None (default): Desktop device
- "mobile": Generic mobile device
- "ios": iOS device (iPhone)
- "android": Android device
parse_results
(optional, bool): Whether to return parsed JSON results. Default is False, which returns raw HTML response.
Invocationโ
Basic Usageโ
from langchain_brightdata import BrightDataSERP
# Initialize the tool
serp_tool = BrightDataSERP(
bright_data_api_key="your-api-key" # Optional if set in environment variables
)
# Run a basic search
results = serp_tool.invoke("latest AI research papers")
print(results)
Advanced Usage with Parametersโ
from langchain_brightdata import BrightDataSERP
# Initialize with default parameters
serp_tool = BrightDataSERP(
bright_data_api_key="your-api-key",
search_engine="google", # Default
country="us", # Default
language="en", # Default
results_count=10, # Default
parse_results=True, # Get structured JSON results
)
# Use with specific parameters for this search
results = serp_tool.invoke(
{
"query": "best electric vehicles",
"country": "de", # Get results as if searching from Germany
"language": "de", # Get results in German
"search_type": "shop", # Get shopping results
"device_type": "mobile", # Simulate a mobile device
"results_count": 15,
}
)
print(results)
Customization Optionsโ
The BrightDataSERP tool accepts several parameters for customization:
Parameter | Type | Description |
---|---|---|
query | str | The search query to perform |
search_engine | str | Search engine to use (default: "google") |
country | str | Two-letter country code for localized results (default: "us") |
language | str | Two-letter language code (default: "en") |
results_count | int | Number of results to return (default: 10) |
search_type | str | Type of search: None (web), "isch" (images), "shop", "nws" (news), "jobs" |
device_type | str | Device type: None (desktop), "mobile", "ios", "android" |
parse_results | bool | Whether to return structured JSON (default: False) |
Use within an agentโ
from langchain_brightdata import BrightDataSERP
from langchain_google_genai import ChatGoogleGenerativeAI
from langgraph.prebuilt import create_react_agent
# Initialize the LLM
llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash", google_api_key="your-api-key")
# Initialize the Bright Data SERP tool
serp_tool = BrightDataSERP(
bright_data_api_key="your-api-key",
search_engine="google",
country="us",
language="en",
results_count=10,
parse_results=True,
)
# Create the agent
agent = create_react_agent(llm, [serp_tool])
# Provide a user query
user_input = "Search for 'best electric vehicles' shopping results in Germany in German using mobile."
# Stream the agent's output step-by-step
for step in agent.stream(
{"messages": user_input},
stream_mode="values",
):
step["messages"][-1].pretty_print()
API referenceโ
Relatedโ
- Tool conceptual guide
- Tool how-to guides