langchain-zenrows

Note: This is a fork of the original langchain-zenrows package, extended to add support for Adaptive Stealth Mode.

The langchain-zenrows integration tool enables LangChain agents to scrape and access web content at any scale using ZenRows' enterprise-grade infrastructure.

Whether you need to scrape JavaScript-heavy single-page applications, bypass anti-bot systems, access geo-restricted content, or extract structured data at scale, this integration provides the tools and reliability needed for modern AI applications.

Table of Contents

• Installation
• Usage
• API Reference
• Features
• License

Installation

pip install langchain-zenrows-adaptive

To use the ZenRows Universal Scraper with LangChain, you'll need a ZenRows API key. You can sign up for free at ZenRows.

For more comprehensive examples and use cases, see the examples/ folder.

Basic Usage

import os
from langchain_zenrows import ZenRowsUniversalScraper

# Set your ZenRows API key
os.environ["ZENROWS_API_KEY"] = "<YOUR_ZENROWS_API_KEY>"

# Initialize the tool
scraper = ZenRowsUniversalScraper()

# Scrape a simple webpage
result = scraper.invoke({"url": "https://httpbin.io/html"})
print(result)

Advanced Usage with Parameters

import os
from langchain_zenrows import ZenRowsUniversalScraper

# Set your ZenRows API key
os.environ["ZENROWS_API_KEY"] = "<YOUR_ZENROWS_API_KEY>"

scraper = ZenRowsUniversalScraper()

# Scrape with JavaScript rendering and premium proxies
result = scraper.invoke({
    "url": "https://www.scrapingcourse.com/ecommerce/",
    "js_render": True,
    "premium_proxy": True,
    "proxy_country": "us",
    "response_type": "markdown",
    "wait": 2000  # Wait 2 seconds after page load
})

print(result)

See the API Reference section below for more available parameters and customizing scraping requests.

Adaptive Stealth Mode

Let ZenRows automatically pick the cheapest configuration that succeeds for each request — escalating to JavaScript rendering or Premium Proxies only when needed. You are billed only for what actually works.

import os
from langchain_zenrows import ZenRowsUniversalScraper

os.environ["ZENROWS_API_KEY"] = "<YOUR_ZENROWS_API_KEY>"

scraper = ZenRowsUniversalScraper()

# ZenRows will automatically determine the best scraping strategy
result = scraper.invoke({
    "url": "https://www.scrapingcourse.com/ecommerce/",
    "mode": "auto",
    "response_type": "markdown",
})

print(result)

Note: When mode is set to "auto", the js_render and premium_proxy parameters are managed by ZenRows automatically and should not be set manually. You can still combine mode="auto" with proxy_country, js_instructions, and custom_headers.

Using with LangChain Agents

from langchain_zenrows import ZenRowsUniversalScraper
from langchain_openai import ChatOpenAI  # or your preferred LLM
from langgraph.prebuilt import create_react_agent
import os

# Set your ZenRows API key
os.environ["ZENROWS_API_KEY"] = "<YOUR_ZENROWS_API_KEY>"
os.environ["OPENAI_API_KEY"] = "<YOUR_OPEN_AI_API_KEY>"


# Initialize components
llm = ChatOpenAI(model="gpt-4o-mini")
zenrows_tool = ZenRowsUniversalScraper()

# Create agent
agent = create_react_agent(llm, [zenrows_tool])

# Use the agent
result = agent.invoke(
    {
        "messages": "Scrape https://news.ycombinator.com/ and list the top 3 stories with title, points, comments, username, and time."
    }
)

print("Agent Response:")
for message in result["messages"]:
    print(f"{message.content}")

CSS Extraction

Extract specific data using CSS selectors:

import json
import os
from langchain_zenrows import ZenRowsUniversalScraper

# Set your ZenRows API key
os.environ["ZENROWS_API_KEY"] = "<YOUR_ZENROWS_API_KEY>"

scraper = ZenRowsUniversalScraper()

# Extract specific elements
css_selector = json.dumps({
    "title": "h1",
    "paragraphs": "p"
})

result = scraper.invoke({
    "url": "https://httpbin.io/html",
    "css_extractor": css_selector
})

Premium Proxy with Geo-targeting

Access geo-restricted content:

import os
from langchain_zenrows import ZenRowsUniversalScraper

# Set your ZenRows API key
os.environ["ZENROWS_API_KEY"] = "<YOUR_ZENROWS_API_KEY>"

scraper = ZenRowsUniversalScraper()

# Check your IP location
result = scraper.invoke({
    "url": "https://httpbin.io/ip",
    "premium_proxy": True,
    "proxy_country": "us"
})
print(result)  # Shows the US IP being used

API Reference

ZenRowsUniversalScraper

Main tool class for web scraping with ZenRows.

Parameters:

• zenrows_api_key (str, optional): Your ZenRows API key. If not provided, looks for ZENROWS_API_KEY environment variable.

Input Schema:

For complete parameter documentation and details, see the official ZenRows API Reference.

Features

• Adaptive Stealth Mode: Automatically selects the optimal scraping configuration, billing only for what succeeds
• JavaScript Rendering: Scrape modern SPAs and dynamic content
• Anti-Bot Bypass: Bypass sophisticated bot detection systems
• Geo-Targeting: Access region-specific content with 190+ countries
• Multiple Output Formats: HTML, Markdown, Plaintext, PDF, Screenshots
• CSS Extraction: Target specific data with CSS selectors
• Structured Data Extraction: Automatically extract emails, phone numbers, links, and other data types
• Session Management: Maintain consistent sessions across requests
• Wait Conditions: Smart waiting for dynamic content
• Premium Proxies: 55M+ residential IPs for maximum success rates

License

langchain-zenrows is distributed under the terms of the MIT license.

Support

• ZenRows Documentation
• LangChain Documentation