HasData MCP Server
Model Context Protocol server for HasData scraping and search APIs. Connect any MCP-compatible AI client to ready-to-use data tools across search, e-commerce, maps, travel, real estate, jobs and social, or pick a single site from the standalone servers.
Quick Start
Two ways to authenticate:
- OAuth (recommended). Connect the server and sign in with your HasData account in the browser — no key to copy around.
- API key. Send your key in the
x-api-keyheader. Get it from the HasData dashboard.
Claude Desktop / Claude.ai. Go to Settings → Connectors → Add custom connector and enter:
https://mcp.hasdata.com/mcp
You'll be prompted to sign in with OAuth. Prefer an API key? Add to claude_desktop_config.json instead:
{
"mcpServers": {
"hasdata": {
"type": "http",
"url": "https://mcp.hasdata.com/mcp",
"headers": {
"x-api-key": "<your-api-key>"
}
}
}
}
Claude Code. With OAuth:
claude mcp add --transport http hasdata https://mcp.hasdata.com/mcp
claude mcp login hasdata
(or run /mcp inside a session and pick Authenticate). With an API key:
claude mcp add --transport http hasdata https://mcp.hasdata.com/mcp --header "x-api-key: <your-api-key>"
Cursor. Add to ~/.cursor/mcp.json or .cursor/mcp.json:
{
"mcpServers": {
"hasdata": {
"url": "https://mcp.hasdata.com/mcp",
"headers": {
"x-api-key": "<your-api-key>"
}
}
}
}
Cline. Add the hasdata server to your Cline MCP settings file:
{
"mcpServers": {
"hasdata": {
"url": "https://mcp.hasdata.com/mcp",
"type": "streamableHttp",
"headers": {
"x-api-key": "<your-api-key>"
},
"disabled": false
}
}
}
Any other MCP client that supports streamable HTTP:
Requests without OAuth or a valid key return 401 Unauthorized.
Need fewer tools in your client? Limit the exposed APIs with the apis query parameter, e.g. https://mcp.hasdata.com/mcp?apis=amazon,google_maps. The apis key for every site is listed in One site per server.
What you can do with it
"Find the cheapest flights from NYC to London next month and compare prices."
"Monitor iPhone 16 prices on Amazon across the last 7 days."
"Pull all 5-star reviews for this product ASIN and summarize the common themes."
"Get me all software engineer job postings in Berlin from Indeed and Glassdoor."
"Find the top-rated Italian restaurants in Chicago using Google Maps."
"What are people searching for related to 'AI agents' on Google Trends this week?"
No scraping code. No proxies to manage. No parsing logic. Just ask.
Compatible clients
Any client that supports the MCP streamable HTTP transport will work.
Tools
Tools across search, e-commerce, maps, travel, real estate, jobs, social, and more. Every group below also ships as standalone servers, one per site, with the same tools behind a shorter URL.
Web
Standalone server for this group: web-scraping-mcp.
Standalone servers for this group: google-search-mcp (all google_serp_* tools), google-images-mcp, google-maps-mcp, google-scholar-mcp, google-trends-mcp, google-flights-mcp, google-flights-deals-mcp, google-hotels-mcp.
Other search engines
Standalone servers for this group: bing-mcp, duckduckgo-mcp.
AI answer engines
Standalone server for this group: chatgpt-mcp.
E-commerce
Standalone servers for this group: amazon-mcp, walmart-mcp, shopify-mcp.
Real Estate
Standalone servers for this group: zillow-mcp, redfin-mcp.
Jobs
Standalone servers for this group: indeed-mcp, glassdoor-mcp.
Travel & Local
Standalone servers for this group: airbnb-mcp, booking-mcp, yelp-mcp, yellowpages-mcp.
Social & Video
Standalone servers for this group: facebook-mcp, instagram-mcp, tiktok-mcp, youtube-mcp.
Prompts and resources
The server ships 20 prompts, ready-made workflows a client can offer instead of making the user compose a tool call.
Alongside them the server exposes 151 resources, one per parameter whose accepted values are a fixed list, across 28 services. Reading one is cheaper than learning the vocabulary from a rejected call, and it costs no credits. Each URI is hasdata://<service>/<parameter>, and resources/list returns the current set.
Both lists are served without an API key, so a client can read them before a user has signed up.
One site per server
If your agent only needs one site, use its standalone server instead of the full set. Each one is the same hosted endpoint filtered with the apis parameter, so the tools, the schemas and the billing are identical, and the model sees a short tool list. Every standalone server also ships as an npm and a PyPI package for clients that want a local stdio launcher, and is listed in the official MCP registry under com.hasdata/<name>.
The remote URL for any of them is https://mcp.hasdata.com/mcp?apis=<key>, and keys can be combined with commas.
Why HasData
A lot of scraping APIs hand you back raw HTML and call it a day. With JavaScript-heavy sites, half of them just time out. HasData's tools return parsed JSON specific to each site, so the model gets structured fields it can actually work with, not a blob of markup to wade through.
A few things worth knowing before you start:
- Each tool is site-specific.
amazon_productknows about ASINs, variants, and seller data.google_maps_placeknows about hours, coordinates, and ratings. You're not calling a generic scraper and hoping for the best. - It's a remote server. Nothing runs locally. No Node process, no version pinning, no cold starts on your machine.
- Credits work the same as the REST API. If you're already a HasData customer, your existing balance and rate limits apply here too.
- Two ways to authenticate. Sign in once with OAuth, or send the
x-api-keyheader on every call — both work against the same endpoint.
Use cases
Competitive intelligence. Point amazon_search and google_serp_shopping at the same query and let the model compare pricing, ranking, and review counts across both. Runs in one prompt.
Lead generation. google_maps_search by keyword and city returns business names, addresses, phone numbers, and ratings. Stack it with yellowpages_search for broader coverage.
Market research. google_trends_search plus google_serp_news in the same session gives you both the search volume signal and the editorial context behind it.
Real estate. zillow_listing accepts filters for beds, baths, price range, and listing type. The model can pull a filtered list and summarize it without you touching a browser.
Recruiting. indeed_listing and glassdoor_listing both take location and keyword. Run both and the model can deduplicate and rank by recency.
Travel. google_travel_flights returns structured itineraries with prices, stops, and duration. Combine it with airbnb_listing for a full trip plan in one conversation.
Billing
Each tool call uses HasData credits the same way a direct API call would. See Credits and Concurrency for details.