MCP Catalogs
Home

cortex-scout

by cortex-works·65·Score 46

Unified web extraction and stateful automation engine for AI agents with MCP integration.

browser-automationweb-scrapingai-llm
7
Forks
0
Open issues
1 mo ago
Last commit
2d ago
Indexed

Overview

CortexScout is a Rust-based MCP server that provides comprehensive web search, extraction, and browser automation capabilities for AI agents. It offers stateful browser automation through a single omni-tool, anti-bot handling with progressive fallbacks, and optional Human-in-the-Loop support. The server exposes a variety of MCP tools for search, crawling, extraction, and deep research, making it suitable for complex web scraping and automation tasks.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

you:AI agents requiring token-efficient web retrieval with anti-bot handling
you:Automated testing workflows replacing heavy frameworks like Playwright
you:Research systems needing deep multi-hop search and extraction
you:How is this different from Playwright MCP?
you:What anti-bot protections does it handle?

When to choose this

Choose CortexScout when your AI agents need reliable web scraping and stateful browser automation with anti-bot protection, especially when working with complex authentication flows or JavaScript-heavy sites.

When NOT to choose this

Avoid if you need open-source licensing (CortexScout uses MIT license but is part of a commercial ecosystem), or if your use case requires only simple static page scraping without browser automation.

Tools this server exposes

9 tools extracted from the README
  • web_search

    URL discovery or search+content retrieval

  • web_fetch

    Unified fetch family (single/batch/crawl modes)

  • extract_fields

    Primary structured extraction from web content

  • scout_browser_automate

    Stateful omni-tool for browser automation

  • memory_search

    LanceDB-backed research history search

  • deep_research

    Multi-hop search + scrape + synthesis

  • scout_agent_profile_auth

    Launch visible browser profile for auth challenges

  • visual_scout

    Visual analysis of web content

  • hitl_web_fetch

    Human-in-the-loop web fetch with auth modes

Comparable tools

browserscapescrapingbeeapifyfirecrawlplaywright-mcp

Installation

Installation

Option A — Prebuilt binaries

Download the latest release assets from GitHub Releases:

  • cortex-scout-mcp — MCP stdio server (recommended for VS Code/Cursor/Claude Desktop)
  • cortex-scout — optional HTTP server

Option B — Build from source

Install protoc first, then:

git clone https://github.com/cortex-works/cortex-scout.git
cd cortex-scout
cargo build --release --manifest-path mcp-server/Cargo.toml --bin cortex-scout-mcp

Claude Desktop Integration

Add to your Claude Desktop config:

{
  "mcpServers": {
    "cortex-scout": {
      "command": "env",
      "args": [
        "RUST_LOG=warn",
        "CORTEX_SCOUT_TOOL_TIMEOUT_SECS=90",
        "/path/to/cortex-scout-mcp"
      ]
    }
  }
}

FAQ

How is this different from Playwright MCP?
CortexScout provides a stateful omni-tool that combines multiple Playwright actions into a single call, reducing token usage for AI agents. It also includes anti-bot handling and optional Human-in-the-Loop capabilities.
What anti-bot protections does it handle?
CortexScout handles various protections through progressive fallbacks: Cloudflare, CAPTCHA, reCAPTCHA, Turnstile, DataDome, AWS Shield, and custom auth challenges. It includes validation evidence against targets like LinkedIn, Ticketmaster, and Amazon.

Compare cortex-scout with

GitHub →

Last updated · Auto-generated from public README + GitHub signals.