Zenrows vs Browserbase Benchmark

This project benchmarks Zenrows against Browserbase for accessing protected and dynamic web pages.

The benchmark, conducted on the 24th of September 2026, sends repeated requests to the same five websites using both platforms and measures successful page retrieval, response times, and HTTP status codes.

Zenrows is tested using mode=auto, while Browserbase is tested using its default Fetch configuration.

Features

• Benchmarks Zenrows and Browserbase against the same five targets
• Sends 100 requests per target
• Uses a controlled rate of 2 requests per second
• Tests different types of web pages
• Validates returned pages using expected content
• Records API and target HTTP status codes
• Measures response time for every request
• Calculates success rates
• Saves raw benchmark results as CSV files
• Supports running either platform independently or both platforms

Prerequisites

Before running the benchmark, make sure you have:

• Python 3.9 or later
• A Zenrows API key (create a free account)
• A Browserbase API key

Installation

Clone the repository:

git clone https://github.com/ZenRows/zenrows-vs-browserbase-benchmark.git
cd zenrows-vs-browserbase-benchmark

Install the dependencies:

pip install -r requirements.txt

Configuration

Copy the example environment file:

cp .env.example .env

Add your API keys to .env:

ZENROWS_API_KEY=your-zenrows-api-key
BROWSERBASE_API_KEY=your-browserbase-api-key

The benchmark loads these environment variables when the script runs. The .env file is excluded from Git through .gitignore, so your API keys are not committed to the repository.

Project structure

.
├── results
│   ├── browserbase_results.csv
│   ├── summary_browserbase.csv
│   ├── summary_zenrows.csv
│   └── zenrows_results.csv
├── .env.example
├── .gitignore
├── LICENSE
├── README.md
├── requirements.txt
└── run_benchmark.py

• run_benchmark.py contains the benchmark script.
• results/zenrows_results.csv contains the raw Zenrows benchmark results.
• results/browserbase_results.csv contains the raw Browserbase benchmark results.
• results/summary_zenrows.csv contains the summarized Zenrows benchmark results.
• results/summary_browserbase.csv contains the summarized Browserbase benchmark results.
• .env.example provides the environment variable names required by the benchmark.
• .gitignore excludes local environment files and other files that should not be committed.
• requirements.txt lists the dependencies required to run the benchmark.
• LICENSE is the MIT license the project is released under.

How it works

The benchmark uses the same five target pages for both platforms and sends 100 requests to each target at a controlled rate of 2 requests per second.

For each request, the script records:

• Target URL
• Platform
• API HTTP status code
• Target HTTP status code
• Expected content match
• Success status
• Response time
• Timestamp

A request is considered successful when the returned page meets the expected content checks defined for that target.

The benchmark runs Zenrows with mode=auto and Browserbase with its default Fetch configuration. Each platform can also be run independently.

Running the project

Run the benchmark:

python run_benchmark.py

The benchmark can also run either platform independently:

python run_benchmark.py --tool zenrows

python run_benchmark.py --tool browserbase

Or run both platforms:

python run_benchmark.py --tool both

Make sure your Zenrows and Browserbase API keys are configured in .env before starting the script.

Output

The benchmark produces four CSV files in the results/ directory:

• zenrows_results.csv contains request-level results for Zenrows.
• browserbase_results.csv contains request-level results for Browserbase.
• summary_zenrows.csv contains summarized results for Zenrows.
• summary_browserbase.csv contains summarized results for Browserbase.

The raw results can be used to inspect individual requests, HTTP status codes, response times, and success status. The summary files provide aggregated benchmark results for each target.

Summary file columns:

Column	Meaning
tool	zenrows or browserbase
target	Target name, matching TARGETS in the script
url	Target URL
total	Requests sent
successful / failed	Request counts
success_rate_pct	Percentage successful
avg_ms, p50_ms, p95_ms	Response time in milliseconds

A request counts as successful only when the API returns 200, the response has a page title, and the page contains that target's phrases from EXPECTED_CONTENT. A target with no EXPECTED_CONTENT entry scores 0%, so the script now refuses to start if one is missing.

Technologies

• Python
• Zenrows Fetch API
• Browserbase Fetch API (called over REST, no SDK)
• aiohttp
• BeautifulSoup + lxml
• python-dotenv
• CSV

Related article

This benchmark is part of the blog:

Best Browserbase Alternative in 2026 for Protected Web Access