Automate the web with Browser MCP
Navigate websites, interact with dynamic pages, run JavaScript, handle multi-step browser workflows, and work with protected websites through the 2Captcha MCP Server.
Browser MCP: Browser automation for AI agents
- Automate clicks, forms, and navigation
- Built-in captcha solving and anti-bot handling
- Proxy support for browser sessions
Browser MCP server gives AI agents access to a managed headless browser for navigating websites, interacting with dynamic pages, filling forms, clicking elements, and extracting data through MCP.
Your agent can launch browser sessions automatically when direct page access is not enough, with built-in proxy, captcha solving, and anti-bot support.
- Headless browser
- Browser automation
- Works with any MCP client
- Anti-bot handling
- Proxy support
The best customer experience provided by leading Browser MCP server
How Browser MCP Server works
Give your AI agent a browser task. 2Captcha Browser MCP Server launches a managed browser, handles website access, and returns the result.
- Give your agent a task
Ask your AI to open a website, navigate pages, click elements, fill forms, or extract data.
- Launch a browser session
Browser MCP Server starts a managed headless browser session for the requested task.
- Interact with the website
The agent navigates pages, executes JavaScript, clicks elements, fills forms, and works with dynamic content.
- Handle protection
Proxy access, anti-bot checks, browser fingerprints, and captcha challenges are handled when required.
- Get the result
Your agent receives page content, extracted data, browser output, or task results and continues the workflow.
Connect Browser MCP
Connect Browser MCP through the 2Captcha MCP Server using the hosted server or run the official package locally.
Hosted MCP: Connect directly to the hosted 2Captcha MCP Server without installing or running your own server.
Local MCP: Run the official @2captcha/mcp package locally and enable the browser tool group.
MCP clients: Works with Claude Code, Claude Desktop, Cursor, Codex, VS Code, and other MCP-compatible clients.
Connect directly to the hosted 2Captcha MCP Server using your API key.
URL: https://mcp.2captcha.com/mcp
Authorization: Bearer YOUR_API_TOKENConnect Claude Code directly to the hosted 2Captcha MCP Server.
claude mcp add --transport http 2captcha https://mcp.2captcha.com/mcp \
--header "Authorization: Bearer YOUR_API_TOKEN"Run the official @2captcha/mcp package locally with npx.
{
"mcpServers": {
"2captcha": {
"command": "npx",
"args": ["@2captcha/mcp"],
"env": {
"API_TOKEN": "YOUR_API_TOKEN",
"GROUPS": "browser"
}
}
}
}Run the official @2captcha/mcp package locally with npx.
{
"mcpServers": {
"2captcha": {
"command": "npx",
"args": ["@2captcha/mcp"],
"env": {
"API_TOKEN": "YOUR_API_TOKEN",
"GROUPS": "browser"
}
}
}
}Run the official @2captcha/mcp package locally with npx.
{
"servers": {
"2captcha": {
"command": "npx",
"args": ["@2captcha/mcp"],
"env": {
"API_TOKEN": "YOUR_API_TOKEN",
"GROUPS": "browser"
}
}
}
}Automate websites in a managed browser
When a task requires real browser interaction, your AI agent can navigate pages, click, type, scroll, run JavaScript, capture screenshots, and work with dynamic content through Browser MCP.
Interact with pages
Navigate, click, type, select options, scroll, hover, and perform multi-step browser actions.
Read and capture content
Get page text or HTML, run JavaScript, capture screenshots, and save pages as PDF files.
Run complex browser workflows
Browser MCP gives AI agents a persistent managed browser session for tasks that require multiple steps, dynamic pages, or authenticated website access.
Maintain browser sessions
Keep cookies and session state while your agent moves between pages and completes multi-step workflows.
Handle protected websites
Use proxy support, browser automation, and captcha solving when websites require additional access handling.
FAQ
What is Browser MCP?
Browser MCP gives AI agents browser automation tools through the 2Captcha MCP Server. Agents can navigate websites, interact with page elements, run JavaScript, capture page content, and manage browser sessions.
Which AI clients can I use with Browser MCP?
It works with Claude Code, Claude Desktop, Cursor, Codex, VS Code, and other MCP-compatible clients.
Do I need to install anything?
No. You can connect to the hosted 2Captcha MCP Server directly. You can also run the official @2captcha/mcp package locally with npx.
How do I enable browser tools?
Browser tools are opt-in. For a local installation, enable them with the browser or browser_full tool group. The browser group provides the main managed-browser tools, while browser_full exposes the complete browser automation toolset.
Can Browser MCP interact with dynamic websites?
Yes. AI agents can navigate pages, click elements, type into fields, select options, scroll, execute JavaScript, and work with JavaScript-rendered content.
Can Browser MCP handle captchas and protected websites?
Yes. 2Captcha MCP includes captcha-solving tools and can use managed browser sessions and proxy access when required by the workflow.
Can Browser MCP keep login sessions?
Yes. Browser tools can work with saved sessions, allowing agents to keep authentication state and reuse it in later browser tasks.
Can I use my own browser instead of the managed browser?
Yes. If you already use Playwright MCP, browser-use, your own Playwright script, or another browser automation tool, you can keep that browser and use 2Captcha MCP specifically for captcha detection and solving.
What is the difference between browser and browser_full?
browser exposes the main browser tools intended for common managed-browser workflows. browser_full exposes the complete Playwright-style browser toolset for more advanced automation.
