Hyperbrowser MCP for AI. Web scraping and browser automation for your AI agent.

Q: How do I scrape data from a page that requires login? (Using startscrape or createsession?)

You must first use createsession to establish a connection, and then execute the necessary authentication steps using runscript. You can then start your scraping job.

Q: Is pagecontent better than extractdata?

pagecontent gives you the entire raw HTML payload, which is useful for auditing. But if you just want names and prices, use extractdata. It's smarter.

Q: What should I do with my sessions when I’m done? (Using stopsession?)

You always call stopsession. This terminates the connection and frees up resources, preventing orphaned or zombie sessions that cost money.

Q: Can I run a script to check if an element exists? (Using runscript?)

Yes. You can pass JavaScript code to runscript that checks for the presence of elements or calculates values on the fly, letting you validate page state.

Q: How do I monitor session health across multiple connections using listsessions?

The listsessions tool provides an overview of all active or past sessions. By filtering the status, you can quickly identify which sessions are running correctly and which ones might be stalled or consuming unnecessary resources.

Claude

ChatGPT

Cursor

Gemini

Windsurf

VS Code

JetBrains

Vercel

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

Connect to your AI in seconds.

Hyperbrowser provides cloud browser infrastructure for your AI agents. It lets you manage remote sessions and scrape dynamic web pages from a natural language prompt.

Use it to extract structured data, execute custom JavaScript on live sites, or capture full-page screenshots without managing any local browser setup.

What your AI can do

Page content

Synchronously retrieves the raw HTML content from a specified URL for immediate processing.

Create session

Starts a new remote browsing session for your agent, allowing you to set proxies or stealth settings.

Extract data

Uses AI intelligence to pull specific structured data (like names and prices) from a rendered webpage URL.

+ 7 more capabilities included

Extract structured data from websites

Your agent uses natural language prompts and optional JSON schemas to pull specific product details or pricing off any live URL.

Manage headless browsing sessions

You can create, list, and terminate remote browser connections, specifying proxies or stealth modes for the job.

Execute custom JavaScript in real time

The agent runs arbitrary JS code inside an active session to interact with page elements or evaluate live page states.

Capture full web pages visually

It takes accurate, high-resolution screenshots of any rendered URL for documentation or visual auditing.

Scrape asynchronous content jobs

You trigger background scraping tasks that handle dynamic loading and retries before retrieving the final HTML payload.

Ask an AI about this

Included with Plan

Waiting for input…

AI Agent

Hyperbrowser (Web Infra for AI) Has 10 Tools

These tools let you start, monitor, and execute actions across complex web pages from your agent.

Make your AI actually useful.

Add this MCP to Claude, Cursor, or Windsurf and your AI stops guessing. It gets real tools to look things up, take action, and handle the stuff you keep doing by hand.

Start using Hyperbrowser (Web Infra for AI) on Vinkius

Page Content

Synchronously retrieves the raw HTML content from a specified URL for immediate processing.

Create Session

Starts a new remote browsing session for your agent, allowing you to set proxies or...

Extract Data

Uses AI intelligence to pull specific structured data (like names and prices) from a...

Get Scrape Job

Checks the status and retrieves the final results of an asynchronous web scraping...

Get Session

Retrieves current details about a specific session, including its duration and...

List Sessions

Provides a list of all active or past browsing sessions, letting you filter by status.

Run Script

Executes custom JavaScript directly inside an active browser session and returns the evaluation result.

Start Scrape

Initiates a background scraping job for a target URL, returning a job ID you can use...

Page Screenshot

Takes and returns metadata or a temporary URL for a full-page screenshot of any...

Stop Session

Terminates an active browsing session immediately, freeing up resources.

Security and governance baked right in.

Pick your AI client below to get set up. Just create a Vinkius account, subscribe, and you're instantly up and running. We handle the entire backend infrastructure, delivering out-of-the-box support for HTTPS Streamable, SSE, and OAuth2—zero messy routing required.

Claude AI

Open Claude Settings

Go to claude.ai, click your profile icon, then navigate to Customize → Connectors.

Add Custom Connector

Click the "+" button and select Add custom connector. Paste your Vinkius endpoint URL:

https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp

Replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com. For OAuth-protected servers, expand Advanced settings to add credentials.

Start a conversation

Open a new chat. The Hyperbrowser integration is available immediately — no restart needed.

Antigravity

Configure Agent Environment

Open your Antigravity agent's workspace configuration or mcp-servers.json file.

Bind the Endpoint

Add the Vinkius endpoint URL to your agent's MCP connections list:

"mcp_servers": {
  "hyperbrowser-web-infra-for-ai": {
    "serverUrl": "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
  }
}

Provide your secure token in place of [YOUR_TOKEN_HERE] to ensure your agent requests are authenticated.

Execute

Start your Antigravity session. The agent will autonomously discover and utilize the Hyperbrowser tools with full Vinkius guardrails applied.

VS Code Copilot

⚡

One-Click Install (Recommended)

In your Vinkius Dashboard, simply click the Add to VS Code button for this server. We'll automatically configure your local workspace.

Or configure manually

Open MCP Settings

Open VS Code, press Ctrl/Cmd + Shift + P, and search for GitHub Copilot: MCP Servers.

Add Server Config

Add the Vinkius endpoint configuration to your mcp-servers.json file:

"hyperbrowser-web-infra-for-ai": {
  "url": "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
}

Ensure you replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com.

LangChain

Install Dependencies

Install the LangChain MCP adapters for your environment:

pip install langchain-mcp-adapters

Connect the Server

Use the SSEClient in LangChain to connect to the Vinkius managed endpoint:

from langchain_mcp_adapters.client import SSEClient

# Connect to Vinkius
client = SSEClient(url="https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp")
tools = client.get_tools()

CrewAI

Define the Tool

Load the Vinkius MCP tools into your CrewAI agents:

from crewai import Agent
from mcp_crewai import MCPTool

# Connect securely to Vinkius
vinkius_tools = MCPTool(url="https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp")

# Assign to Agent
researcher = Agent(
    role='Data Researcher',
    tools=vinkius_tools.get_all()
)

Execute Task

Run your CrewAI process. The agent will autonomously route tasks to the Vinkius managed server.

Choose How to Get Started

Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.

Build Your Own

Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.

Import from OpenAPI, Swagger, or YAML specs
Create Agent Skills with progressive disclosure
Deploy to edge with MCPFusion framework
Built in DLP, auth, and compliance on every call
Real time usage dashboard and cost metering
Publish to catalog or keep private

Start building

Make Your AI Do More

Start with Hyperbrowser (Web Infra for AI), then connect any of our 5,100+ other servers whenever your AI needs more. One click, no limits.

Use this MCP plus 5,100+ others, all in one place
Add new capabilities to your AI anytime you want
Every connection is secured and compliant automatically
Track usage and costs across all your servers
Works with Claude, ChatGPT, Cursor, and more
New servers added to the catalog every week

Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Hyperbrowser. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.

VINKIUS INFRASTRUCTURE

Cloud Hosted

Managed infra

V8 Isolated

Sandboxed per request

Zero-Trust Proxy

No stored credentials

DLP Enforced

Policy on every call

GDPR Compliant

EU data residency

Token Compression

~60% cost reduction

Your data is protected. See how we built it.

Works with Claude, ChatGPT, Cursor, and more

The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.

This connection provides 10 powerful capabilities that interface natively with Claude, ChatGPT, Cursor, and other compatible AI platforms. No middleware. No custom integration required.

Dealing with JavaScript and AJAX hell every day.

Right now, if you need to scrape a pricing page or check product availability on e-commerce sites, the workflow is brutal. You open the URL, wait for it to load, then manually copy text from one section, run another query in a different tab, and stitch everything together in a spreadsheet. If that data depends on a user clicking 'Show More' or waiting 3 seconds for an API call, your current process breaks every time.

With this MCP, you tell the agent what to do—like, 'Click the pricing tier and extract the monthly cost.' The underlying infrastructure handles the clicks, the wait times, and the JavaScript execution. You just get clean, structured data back.

The `extract_data` tool gives you reliable structure.

Previously, pulling product details meant writing brittle scrapers that relied on specific CSS class names. When the target site updated its design—even changing one wrapper ID—your entire script broke. You'd spend half a day debugging selectors instead of analyzing data.

Now you use `extract_data`. You give your agent a natural language request and maybe a simple JSON schema. The MCP handles the web rendering, figuring out where the 'product name' actually lives, regardless of how messy the site's underlying code is.

Support 24/7 support@vinkius.com ↗

Security Vinkius Trust Center ↗

SLA Service Level Agreement ↗

Report Listing Send Report ↗

What your AI can actually do with this

Your agent needs to interact with the real internet—the kind that runs complex JavaScript and loads content dynamically. This MCP gives your AI client full control over cloud web infrastructure. You don't just get raw HTML; you can tell it what data to find, give it a JSON schema, and it returns clean, structured results.

It handles things like proxies, stealth mode, and CAPTCHAs automatically in the background. Because this MCP is hosted on Vinkius, you connect your preferred AI client once and gain access to all these web automation tools. Need to test how a website looks? Capture a full-page screenshot. Want to check if a button works? Run custom JavaScript directly against an active session.

It's basically giving your agent the ability to sit at a powerful remote workstation and browse the web, exactly when you need it.

Built · Hosted · Managed by Vinkius Hyperbrowser-MCP - Web Infra for AI Data Extraction

Server ID 019d75b5-e910-7379-87b2-48d1e15736f2

Vinkius Inspector

Compliance Grade A+

Score 100/100

Report View Report ↗

Here's how it actually works

The bottom line is that you treat web interaction like a managed API call, not messy browser automation.

First, use create_session to establish a new remote browser connection. You can configure proxies or specify stealth settings here.

Next, you either kick off an asynchronous job using start_scrape for large jobs, or run immediate code via run_script against the active session.

Finally, retrieve results by polling the status with get_scrape_job, or pulling raw data instantly using page_content.

Who is this actually for?

Data scientists who need to scrape complex e-commerce sites for market research. QA engineers checking cross-browser compatibility or visual regressions. Ops engineers needing reliable ways to automate web reporting tasks.

Data Scientist

You use the MCP to run extract_data on competitor pricing pages, getting clean JSON output for market analysis without building a dedicated scraping pipeline.

QA Engineer

You repeatedly call page_screenshot with different session configurations (create_session) to verify that button placement hasn't changed across browser versions.

DevOps Engineer

You automate routine web health checks by combining list_sessions, checking status, and running a simple validation script via run_script.

What Changes When You Connect

You don't need to worry about managing local headless browsers. The system handles all the infrastructure, letting you focus purely on the logic of extraction or scripting.

Need a complex multi-step data pull? You can chain tools: use start_scrape first, then monitor completion with get_scrape_job, and finally pass the results to extract_data. It handles state changes automatically.

The ability to run custom JS via run_script means you aren't limited by what standard scrapers can see. You can interact with elements that require clicks or calculations before extraction.

When auditing web apps, calling page_screenshot is a huge time saver. Instead of writing complex visual regression tests, your agent just captures the state and gives you the URL.

Forget boilerplate setup. With this MCP on Vinkius, you connect your preferred AI client once and immediately gain full access to all these web automation tools.

See it in action

01 01

Tracking dynamic pricing changes

A competitor's site loads prices only after a user clicks 'View Details.' Your agent first uses create_session with stealth mode, then executes JavaScript via run_script to click the button. Finally, it calls extract_data on the resulting page view.

02 02

Checking for visual layout bugs

The QA team suspects a banner ad is overlapping content on mobile. Instead of manually testing across devices, they run an agent that uses page_screenshot with specific viewport dimensions to capture and compare the state.

03 03

Collecting structured data from multiple pages

A research task requires grabbing product names and specs from 50 different category URLs. The agent runs a bulk scrape using start_scrape for all links, then uses get_scrape_job to collect the resulting payloads before running extract_data on the full set.

04 04

Validating API endpoints through UI

The team needs to confirm that a specific form submission works correctly. They use an agent to navigate to the page, then call run_script to simulate the user interaction and check the returned value.

The honest tradeoffs

Assuming data is always present

Anti-pattern

Just calling page_content on a dynamic page assumes all necessary scripts ran, giving you incomplete or broken HTML that can't be parsed.

The Fix

Always pair create_session with an active interaction. Use run_script first to force the required element to load or trigger the AJAX call before calling page_content.

Ignoring session cleanup

Anti-pattern

Running dozens of test jobs without proper shutdown leads to resource leaks, crashing your environment and wasting API credits.

The Fix

Always wrap your workflow: after the job is done or data is extracted, call stop_session immediately. If you forget, use list_sessions first.

Over-relying on raw scraping

Anti-pattern

Getting a large payload from start_scrape and trying to write regex against it is slow, brittle, and fails when the site changes one tag name.

The Fix

Use the dedicated extract_data tool. You provide the schema; the AI handles the messy parsing logic for you.

Questions you might have

How do I scrape data from a page that requires login? (Using `start_scrape` or `create_session`?) +

You must first use create_session to establish a connection, and then execute the necessary authentication steps using run_script. You can then start your scraping job.

Is `page_content` better than `extract_data`? +

page_content gives you the entire raw HTML payload, which is useful for auditing. But if you just want names and prices, use extract_data. It's smarter.

What should I do with my sessions when I’m done? (Using `stop_session`?) +

You always call stop_session. This terminates the connection and frees up resources, preventing orphaned or zombie sessions that cost money.

Can I run a script to check if an element exists? (Using `run_script`?) +

Yes. You can pass JavaScript code to run_script that checks for the presence of elements or calculates values on the fly, letting you validate page state.

When I use `start_scrape`, and the job status isn't 'completed', what should I check using `get_scrape_job`? +

You need to examine the response payload for error codes or metadata. If the job failed, the result will contain a clear failure reason—it could be rate limiting, an invalid URL, or a CAPTCHA block. You can then adjust your configuration and resubmit.

When creating a session, how does `create_session` handle advanced proxy configurations and stealth modes? +

You pass all specific details—like rotation groups or desired browser versions—in the optional JSON config. This lets you manage complex network requirements in one call, ensuring your agent avoids detection and maintains connection stability.

If I execute custom JS using `run_script`, can I then use `page_content` to capture the updated DOM structure? +

Yes. You run the script first to manipulate the page state, and immediately follow up by calling page_content. This guarantees that the raw HTML payload reflects the changes you just made, giving you a highly accurate snapshot of the current web element tree.

How do I monitor session health across multiple connections using `list_sessions`? +

The list_sessions tool provides an overview of all active or past sessions. By filtering the status, you can quickly identify which sessions are running correctly and which ones might be stalled or consuming unnecessary resources.

Can my agent extract specific product data from an e-commerce site? +

Yes. Use the extract_data tool and provide the target URL along with a prompt like 'Extract product name, price, and customer rating'. Hyperbrowser's LLM will render the page and return clean JSON data according to your request.

How do I handle websites that require dynamic JavaScript loading? +

Hyperbrowser is built for dynamic content. Tools like page_content and start_scrape wait for JavaScript to settle before returning results, ensuring your agent sees the fully rendered version of any web application.

Can I run custom automation scripts in the cloud? +

Absolutely. The run_script tool allows you to execute JavaScript strings inside an active browser session. This is perfect for simulating complex user interactions or extracting data that requires custom logic.

Connect to your AI in seconds.

Page content

Create session

Extract data

Hyperbrowser (Web Infra for AI) Has 10 Tools

Make your AI actually useful.

Page Content

Create Session

Extract Data

Get Scrape Job

Get Session

List Sessions

Run Script

Start Scrape

Page Screenshot

Stop Session

Security and governance baked right in.

Claude AI

Open Claude Settings

Add Custom Connector

Start a conversation

Claude Code

Open your terminal

Add the MCP Server

Start coding

Cursor

One-Click Install (Recommended)

Open Cursor Settings

Add New Server

Use in Composer

Antigravity

Configure Agent Environment

Bind the Endpoint

Execute

VS Code Copilot

One-Click Install (Recommended)

Open MCP Settings

Add Server Config

Windsurf

One-Click Install (Recommended)

Open Windsurf Settings

Add Server Endpoint

LangChain

Install Dependencies

Connect the Server

CrewAI

Define the Tool

Execute Task

Choose How to Get Started

Build Your Own

Make Your AI Do More

Works with Claude, ChatGPT, Cursor, and more

Dealing with JavaScript and AJAX hell every day.

The `extract_data` tool gives you reliable structure.

What your AI can actually do with this

Here's how it actually works

Who is this actually for?

What Changes When You Connect

See it in action

Tracking dynamic pricing changes

Checking for visual layout bugs

Collecting structured data from multiple pages

Validating API endpoints through UI

The honest tradeoffs

Assuming data is always present

Ignoring session cleanup

Over-relying on raw scraping

When It Fits, When It Doesn't

Questions you might have