Let your AI assistant scrape. Claude · ChatGPT · Cursor.
Add one URL and your assistant can find pages, extract data, run your scrapers and read the results from a chat. You sign in with your Simplescraper account, so there is no API key to paste.
Connected to mcp.simplescraper.io/mcp{ "domain": "https://acmeoutdoor.com" }214 URLs from https://acmeoutdoor.com/sitemap.xml
{ "recipe_id": "Qm7xT2bL9a", "batch_mode": "replace", "batch_urls": [ 38 boot pages ] }success · 38 new URLs · 0 errors
{ "recipe_id": "Qm7xT2bL9a", "use_crawler": true }status completed · 38 rows
| Boot | Price | Rating |
|---|---|---|
| Basecamp Chukka | $99.00 | 4.4 |
| Canyon Mid | $118.00 | 4.5 |
| Ridge Hiker Low | $129.00 | 4.6 |
| Trailrunner GTX | $139.00 | 4.7 |
| Summit Approach | $145.00 | 4.5 |
extract_urls is free · other tools use your credits
Any MCP client
Claude, ChatGPT, Cursor and other assistants that support MCP connect with the same server URL.
Sign-in without an API key
Your client opens a Simplescraper sign-in page when you connect. The config holds only the server URL.
Free site mapping
Finding a site's URLs costs nothing. The other tools use the same credits as the dashboard.
The 13 MCP tools your assistant can use
Your assistant picks the tool from what you ask. Every tool works on your own account, with the same scrapers, results and credits as the dashboard.
Data extraction
Get data from a page or from a scraper you saved.
smart_extractPulls named fields from any page with AI, with no CSS selectors."Extract product names, prices and ratings from lumensupply.com/packs"extract_urlsFreeFinds every page URL on a site from its sitemap. Uses no credits."Get me a list of all pages on northwind.com"extract_markdownReads one page as clean Markdown, for articles, posts and docs."Read northwind.com/blog/q3-update and summarise it"run_recipeRuns a saved scraper and returns fresh data. Uses credits."Run my competitor price scraper now and show me the new data"get_latest_resultsReturns a scraper's most recent data without a new scrape."Show me the latest results from my job board scraper"get_resultsReturns the data from one run by its results ID."Get the data from results ID abc123"get_results_historyLists a scraper's past runs with dates and page counts."Compare this week's price scraper results with last week's run"Scraper management
Find, read, create and change your scrapers.
list_recipesFinds scrapers by site, name or recent activity."List my scrapers that read from acmeoutdoor.com"get_recipeShows a scraper's full details, including its selectors."Show me the selectors in my price tracker"create_recipeBuilds a new scraper from CSS selectors."Create a scraper for brightline.ai/blog that grabs titles and dates"update_recipeChanges a scraper's name, URL or selectors."Change the price selector on my boot scraper to .price-now"Batch
Give a scraper a list of pages to run through.
update_batch_urlsAdds to or replaces the URLs a scraper crawls. Pairs with extract_urls and run_recipe."Add these 10 product URLs to my product details scraper"Screenshots
See a page the way a browser shows it.
screenshotCaptures a page and returns a hosted image link. 1 credit each."Take a full-page screenshot of parallaxlabs.com"Connect any MCP client with one URL
Every client takes the same server URL. When you connect, the client opens a Simplescraper sign-in page and the connection works only on your account.
Your account speaks MCP.
Paste the URL into any client that supports remote MCP servers. In Claude Code it is one command.
Claude Desktop
- Open Settings and go to Integrations.
- Click Add MCP Server.
- Name it Simplescraper and paste the server URL.
- Sign in to connect your Simplescraper account, then ask in plain words.
Cursor
{
"mcpServers": {
"simplescraper": {
"url": "https://mcp.simplescraper.io/mcp"
}
}
}Cursor opens your browser to sign in. Leave out any Authorization header, because it stops the sign-in from working. Put the file in a project's .cursor folder to use it in that project only.
How to set up the MCP server
Once connected, your assistant uses Simplescraper whenever a request needs data from the web or from your scrapers.
Add the server URL
Paste it into Claude, ChatGPT, Cursor or any client that supports MCP.
Sign in with Simplescraper
Your client opens a sign-in page. No API key goes into the config.
Ask in plain words
Find pages, extract fields, run scrapers and compare results from a chat.
FAQ
What is the Simplescraper MCP server?
MCP (Model Context Protocol) lets AI assistants like Claude Desktop and Cursor directly control Simplescraper. Connect once, then use natural language to create scrapers, run scrapes, and analyze data. Just say "extract all products from this website" and your AI handles the rest.
How do I connect Claude or Cursor to Simplescraper?
For Claude Desktop: Go to Settings → Integrations → Add MCP Server, then enter our URL: https://mcp.simplescraper.io/mcp. For Cursor: Add the same URL to your mcp.json config file. After authenticating, you can control Simplescraper with natural commands.
What can AI agents do with Simplescraper MCP?
Most of what you do with scrapers in the dashboard. AI can list and search your scrapers, create new scrapers, update selectors, run scrapes, retrieve results, manage batch URLs, take screenshots, and analyze data. Example: "Compare my price tracker's latest results with last week's run and list the items that dropped over 10%."
Can AI agents use Smart Extract without CSS selectors?
Yes! Just tell your AI what to extract in plain English. Say "Extract product names, prices, and ratings from [URL]" and the AI uses Smart Extract to get the data without any technical setup. Perfect for quick data extraction tasks.
Connect your AI assistant to Simplescraper today.
Connect your assistant in a minute, with 100 free credits to start.
Start free