Scrapeunblocker MCP server
{
"mcpServers": {
"scrapeunblocker": {
"command": "npx",
"args": [
"-y",
"scrapeunblocker-mcp@latest"
],
"env": {
"SCRAPEUNBLOCKER_KEY": "<scrapeunblocker_key>"
}
}
}
}{
"mcpServers": {
"scrapeunblocker": {
"command": "npx",
"args": [
"-y",
"scrapeunblocker-mcp@latest"
],
"env": {
"SCRAPEUNBLOCKER_KEY": "<scrapeunblocker_key>"
}
}
}
}{
"mcpServers": {
"scrapeunblocker": {
"command": "npx",
"args": [
"-y",
"scrapeunblocker-mcp@latest"
],
"env": {
"SCRAPEUNBLOCKER_KEY": "<scrapeunblocker_key>"
}
}
}
}{
"servers": {
"scrapeunblocker": {
"command": "npx",
"args": [
"-y",
"scrapeunblocker-mcp@latest"
],
"env": {
"SCRAPEUNBLOCKER_KEY": "${input:scrapeunblocker_key}"
}
}
},
"inputs": [
{
"id": "scrapeunblocker_key",
"type": "promptString",
"description": "Scrapeunblocker key",
"password": true
}
]
}{
"mcpServers": {
"scrapeunblocker": {
"command": "npx",
"args": [
"-y",
"scrapeunblocker-mcp@latest"
],
"env": {
"SCRAPEUNBLOCKER_KEY": "<scrapeunblocker_key>"
}
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsoninstall npx -y scrapeunblocker-mcpversion 0.2.0
What it does
This MCP server lets you fetch HTML from any web page through the ScrapeUnblocker anti-bot API, bypassing protections like Cloudflare and DataDome. It supports interactive browser steps, element discovery, and AI-parsed JSON extraction. Intended for developers who need to scrape protected pages and integrate that data into Claude workflows.
- Fetch HTML from a protected e-commerce product page, interact with it to reveal prices, then extract structured data.
- Discover form fields and buttons on a blocked page using element listing, then automate filling and submitting the form.
- Run Google searches and extract organic results as JSON for research or data collection tasks.
Tools this server exposes
| Tool | Does |
|---|---|
| fetch_html | Fetch the fully rendered HTML of a URL, optionally after running browser interactions like clicks, typing, and scrolling. |
| list_elements | List notable elements on a page (links, buttons, inputs, selects) with ready-to-use selectors and attributes. |
| fetch_parsed | Fetch a page and return AI-parsed structured JSON instead of raw HTML. |
| google_search | Run a Google search and return organic results as JSON. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.