Playwright MCP Server

Give your AI agents a real browser. This MCP server runs Playwright on Apify's infrastructure, letting AI agents navigate websites, fill forms, click buttons, extract data, and take screenshots — all through the Model Context Protocol.

Why Use This?

• No local browser needed — runs on Apify's cloud with anti-blocking proxies
• 16 browser tools — navigate, click, type, screenshot, extract text/HTML, run JavaScript, and more
• Anti-blocking built in — Apify proxy infrastructure handles CAPTCHAs and blocks
• Works with any MCP client — Claude, OpenAI, custom agents via SSE or HTTP transport

Available Tools (16)

Tool	Description
navigate	Go to a URL, wait for page load
screenshot	Capture page or element as PNG
click	Click elements by selector or text
type	Type into input fields
get_text	Extract visible text from page/element
get_html	Get HTML content
evaluate	Run JavaScript in page context
select	Choose dropdown options
scroll	Scroll page or elements
wait	Wait for elements to appear/disappear
get_links	Extract all links with optional filtering
extract_table	Parse HTML tables to JSON
pdf	Generate PDF of current page
fill_form	Fill multiple form fields at once
get_cookies	Read browser cookies
set_cookies	Set browser cookies

Quick Start

Connect via MCP (SSE)

{
  "mcpServers": {
    "playwright": {
      "url": "https://nexgendata--playwright-mcp-server.apify.actor/sse",
      "transport": "sse",
      "headers": {
        "Authorization": "Bearer YOUR_APIFY_API_TOKEN"
      }
    }
  }
}

Connect via HTTP

{
  "mcpServers": {
    "playwright": {
      "url": "https://nexgendata--playwright-mcp-server.apify.actor/mcp",
      "transport": "http"
    }
  }
}

Use Cases

• Web research agents — Browse and extract data from any website
• Form automation — Fill out applications, surveys, registrations
• Testing & monitoring — Check website functionality and content
• Data extraction — Scrape dynamic JavaScript-rendered pages
• Screenshot capture — Visual documentation and monitoring
• Price monitoring — Track prices across e-commerce sites

Example: Navigate and Extract

Agent: "Go to example.com and get all the links"

Tool calls:
1. navigate(url="https://example.com")
2. get_links()

Example: Fill a Form

Agent: "Search for 'web scraping' on Google"

Tool calls:
1. navigate(url="https://google.com")
2. type(selector="textarea[name='q']", text="web scraping", pressEnter=true)
3. wait(selector="#search")
4. get_text(selector="#search")

Proxy Support

Enable Apify proxy for anti-blocking:

{
  "proxyConfiguration": {
    "useApifyProxy": true
  }
}

Pricing

Pay per browser action via Apify's pay-per-event model:

• Session start: $0.05
• Per tool call: $0.01

Support

Questions? Open an issue on the actor page or contact NexGenData.

---

💻 Code Example — Python

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("nexgendata/playwright-mcp-server").call(run_input={
    # Fill in the input shape from the actor's input_schema
})

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item)

🌐 Code Example — cURL

curl -X POST "https://api.apify.com/v2/acts/nexgendata~playwright-mcp-server/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ /* input schema */ }'

❓ FAQ

Q: How do I get started? Sign up at apify.com, grab your API token from Settings → Integrations, and run the actor via the Apify console, API, Python SDK, or any integration (Zapier, Make.com, n8n).

Q: What's the typical cost per run? See the pricing section below. Most runs finish under $0.10 for typical batches.

Q: Is this actor maintained? Yes. NexGenData maintains 165+ Apify actors and ships updates regularly. Bug reports via the Apify console issues tab get responses within 24 hours.

Q: Can I use the output commercially? Yes — you own the output data. Check the target site's Terms of Service for any usage restrictions on the scraped content itself.

Q: How do I handle rate limits? Apify manages concurrency and retries automatically. For very large batches (10K+ items), run multiple smaller jobs in parallel instead of one mega-job for better reliability.

💰 Pricing

Pay-per-event pricing — you only pay for what you actually extract.

• Actor Start: $0.0050
• tool call: $0.0050

🔗 Related NexGenData Actors

• RAG Web Browser
• AI Web Scraper
• Hacker News Scraper

🚀 Apify Affiliate Program

New to Apify? Sign up with our referral link — you get free platform credits on signup, and you help fund the maintenance of this actor fleet.

📚 More From NexGenData

Explore the full catalog, tutorials, Gumroad data packs, and newsletter at thenextgennexus.com — the brand home for everything we ship.

• 📖 Tutorials & how-to guides
• 🗂️ Full actor catalog with usage examples
• 📦 Gumroad data packs (one-time purchases)
• 📬 Newsletter — monthly drops of new actors and revenue experiments

---

Built and maintained by NexGenData — 165+ actors covering scraping, enrichment, MCP servers, and automation. 🏠 Home: thenextgennexus.com

---

Why Playwright MCP Beats Browserbase, Anthropic Computer Use API, Selenium Grid & Generic LLMs

Feature	NexGenData Playwright MCP	Browserbase	Anthropic Computer Use	Selenium Grid (self-host)	Generic LLM (no browser)
Cost	$0.002 / event, pay-per-event	$39+ / month base	$$$ tokens per screenshot	Self-host infra costs	Free (no web access)
MCP-native	Yes	Partial	Native to Claude only	No	No
Headless Chromium / Firefox / WebKit	Yes	Chrome only	Chrome only	Yes	N/A
Apify proxy pool	Yes (residential + DC)	Limited	None	Build it yourself	N/A
Captcha / Cloudflare handling	Yes	Plan-gated	None	Build it yourself	N/A
Screenshot + DOM + console output	All three	Yes	Screenshot only	Yes (manual wiring)	N/A
AI-agent integration	Native MCP — any client	Browserbase SDK	Claude-only	None	N/A
Auth	Apify token	Browserbase key	Anthropic key + tier	None	None
Monthly minimum	None	$39+	Anthropic tier	Infra fixed cost	None

Most teams pick this MCP server because it is cheaper than Browserbase for low-volume agent traffic, works with any MCP client (not just Claude) unlike Anthropic Computer Use, and ships pre-baked Apify proxy + captcha handling that Selenium Grid forces you to build yourself. A drop-in browser tool for any AI agent that needs to load a page, click, fill, and screenshot.

Related NexGenData MCP Servers & Web-Scraping Actors

Use case	Actor
Web scraping MCP (HTTP + parse, no browser)	web-scraping-mcp-server
Developer tools MCP (StackOverflow, GitHub)	developer-tools-mcp-server
Domain intelligence MCP	domain-intelligence-mcp-server
Google Maps MCP	google-maps-mcp-server
News MCP	news-mcp-server
Finance MCP	finance-mcp-server
Real estate MCP	real-estate-mcp-server
Reddit MCP	reddit-mcp-server
YouTube media MCP	youtube-media-mcp-server
Review intelligence MCP	review-intelligence-mcp-server
Academic research MCP	academic-research-mcp-server

Browse the full NexGenData catalog of 260+ actors at https://apify.com/nexgendata?fpr=2ayu9b