# Scrapewise [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/BEBOTECH-OU/scrapewise-mcp  
**GitHub Stars:** 1  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/scrapewise

## Description
Scrape, clean and match product and price data from any website

## Claude Desktop Quick Installation
Remote MCP endpoint (confidence: high). Install path detected from listing signals. Add as a URL/SSE server in your client:

```json
"mcpServers": {
  "scrapewise": {
    "url": "https://mcp.scrapewise.ai/mcp"
  }
}
```

## Documentation

## What Scrapewise MCP server does

Scrapewise MCP server gives an AI agent access to hosted data collection and web research functions. The default endpoint can retrieve product details from shops and Amazon, clean rows that are already available, find businesses in Google Maps, search Google and Amazon, query Google Trends, fetch a public page, and read results saved in a Scrapewise account.

Product records may include names, prices, currencies, availability, brands, SKU or GTIN/EAN identifiers, images, descriptions, and ratings. Amazon product retrieval can also include seller prices, conditions, and delivery information. Local business results include names, addresses, phone numbers, websites, and ratings, with up to 20 results per call.

The optional setup toolset adds scraper construction and administration. Agents can preview pages, create scrapers from URLs or cURL commands, define extracted columns, organize scrapers into groups, run jobs, inspect errors, export data, configure transformations, schedule collection, and manage price monitoring.

## How it works

The service is hosted at `https://mcp.scrapewise.ai/mcp`, so the repository does not contain a local server to install or run. MCP clients connect over Streamable HTTP and send an API key as `Authorization: Bearer <your API key>`.

The bare endpoint exposes the short data-tool list for accounts using a new MCP key. Adding `?toolsets=outcome,admin` requests the setup tools as well. The toolset selection is evaluated when a session starts, so clients must reconnect after changing the URL. What appears can also depend on the API key's access level.

Scrapewise MCP server supports long-running work through progress, result retrieval, and cancellation tools. Paid tools accept `quote_only: true` to return the price and wallet balance without starting paid work. `clean_product_data` operates on existing rows and does not fetch data.

## Setup and configuration

Create an account at Scrapewise, open Settings → API Keys, and generate an API key. Pricing is pay-as-you-go. Configure the remote MCP endpoint in a compatible client with the endpoint URL and Authorization header.

Claude Code uses its HTTP MCP command; Claude Desktop and Cursor use JSON configuration with `url` and `headers`; VS Code uses the MCP user configuration with a `servers` map and `"type": "http"`. Other clients can connect if they support Streamable HTTP. When using the setup URL in a shell, quote the URL because the `?` character can be interpreted by some shells.

## Tools and capabilities

The default tool set includes product retrieval, Amazon product retrieval, data cleaning, local business lookup, Google search, Amazon search, Google Trends, page fetching, job result access, job cancellation, and account data reading.

With the administrative toolset, the Scrapewise MCP server also exposes tools for:

- Building and updating URL-, cURL-, and file-based scrapers
- Creating and managing scraper groups and their product URLs
- Running or stopping individual scrapers, groups, URL lists, and sitemap harvests
- Reading samples, complete result sets, run columns, histories, and errors
- Exporting data and managing result categories
- Applying currency conversion, quantity normalization, and price-per-unit rules
- Configuring schedules, retention, data quality, and price-monitoring competitors

## Limitations and notes

The server is hosted; this repository is a connection manifest and client configuration reference rather than an application to run. Tool availability varies by endpoint and API key. A read-only `LLM_READ` key exposes free reading and cleaning functions, while paid retrieval and scraper-building functions require a full-access `LLM_FULL` key. The documentation states that the default tool list can differ from the setup list and from the tools advertised after connection.

_Full upstream README: https://allmcps.com/mcp/scrapewise/readme_