# MCP Scraper agent retrieval

Query: overview
Retrieval: lexical
Corpus SHA-256: b8e582cf1b92d2e99ab208ffa4cd129c500b7f9a53329837779757ad1137c243
JSON: https://mcpscraper.dev/llms/json

## What MCP Scraper is
https://mcpscraper.dev/
MCP Scraper gives AI agents a shared web-data, memory, integrations, and scheduling layer. It can collect public evidence, keep approved information searchable across tools, connect authorized business accounts, and run bounded recurring work in the cloud.

## Agent surfaces
https://mcpscraper.dev/docs
The product is available through a hosted MCP endpoint, an npm stdio server, a Claude Desktop extension, and a REST API. The same account can be used from Claude, ChatGPT, Cursor, and other compatible agent surfaces.

## Operating boundary
https://mcpscraper.dev/integrations
Public-web collection, private memory, connected-account access, and scheduled actions remain distinct permission lanes. Connected writes require explicit enablement and grants; retrieval results preserve canonical sources and corpus receipts.