Cortex Scout
A unified web extraction and stateful automation engine for AI. Replaces heavy testing frameworks with token-optimized browser control, deep research, and HITL.
What is Cortex Scout?
Cortex Scout is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to unified web extraction and stateful automation engine for ai. replaces heavy testing frameworks with token-optimized browser control, deep research, and hitl.
A unified web extraction and stateful automation engine for AI. Replaces heavy testing frameworks with token-optimized browser control, deep research, and HITL.
This server falls under the Browser Automation and Developer Tools categories on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- A unified web extraction and stateful automation engine for
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx cortex-scoutConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Cortex Scout
Cortex Scout is a unified web extraction and stateful browser automation engine built as an MCP server in Rust, designed to give AI agents token-efficient access to the web. It provides tools for web search across multiple engines, single and batch URL fetching with content filtering, structured field extraction, full browser automation via Playwright, anti-bot bypass, and multi-hop deep research with optional LLM synthesis. Developers integrate it when they need AI agents to reliably retrieve web content, navigate authenticated sites, or conduct deep research without hitting token limits from verbose HTML.
Prerequisites
- A prebuilt cortex-scout-mcp binary from the GitHub releases page, or Rust toolchain and protobuf compiler for building from source
- Chromium, Chrome, or Brave browser installed (auto-detected by the server for automation features)
- An OpenAI-compatible API key (optional, required only for deep research LLM synthesis)
- An MCP-compatible client such as Claude Desktop, Cursor, or VS Code
Download the prebuilt binary
Download the cortex-scout-mcp binary for your platform from the GitHub releases page at github.com/cortex-works/cortex-scout/releases and make it executable.
chmod +x cortex-scout-mcpConfigure your MCP client
Add Cortex Scout to your MCP client configuration. The binary is invoked directly as a stdio server. Set RUST_LOG=warn to avoid flooding stderr with info-level logs.
{
"mcpServers": {
"cortex-scout": {
"type": "stdio",
"command": "env",
"args": [
"RUST_LOG=warn",
"MAX_CONTENT_CHARS=10000",
"SEARCH_ENGINES=google,bing,duckduckgo,brave",
"CORTEX_SCOUT_TOOL_TIMEOUT_SECS=90",
"/path/to/cortex-scout-mcp"
]
}
}
}Enable deep research (optional)
To use the deep_research tool which synthesizes multi-hop search results with an LLM, add your OpenAI-compatible API credentials and enable the feature.
Set up persistent research memory (optional)
Set the LANCEDB_URI environment variable to a directory path to enable persistent vector-based memory across sessions using the memory_search tool.
Restart your MCP client
Restart Claude Desktop, Cursor, or your IDE. Cortex Scout will start as a background stdio process and tools like web_search, web_fetch, extract_fields, and deep_research will become available.
Cortex Scout Examples
Client configuration
Claude Desktop configuration for Cortex Scout with deep research enabled via OpenRouter. Replace the binary path and API key with your actual values.
{
"mcpServers": {
"cortex-scout": {
"type": "stdio",
"command": "env",
"args": [
"RUST_LOG=warn",
"MAX_CONTENT_CHARS=10000",
"SEARCH_ENGINES=google,bing,duckduckgo,brave",
"CORTEX_SCOUT_TOOL_TIMEOUT_SECS=90",
"DEEP_RESEARCH_ENABLED=1",
"OPENAI_API_KEY=sk-or-v1-your-key",
"OPENAI_BASE_URL=https://openrouter.ai/api/v1",
"DEEP_RESEARCH_LLM_MODEL=gpt-4o-mini",
"/path/to/cortex-scout-mcp"
]
}
}
}Prompts to try
These prompts demonstrate web search, URL fetching, field extraction, and deep research capabilities.
- "Search the web for the latest news about Rust async runtimes and summarize the top 3 results."
- "Fetch the content of https://example.com and extract only the main article text."
- "Do a deep research report on the current state of MCP adoption in 2025."
- "Automate a login to this site and take a screenshot of the dashboard."
- "Extract the product name, price, and availability from these 5 e-commerce URLs in batch."Troubleshooting Cortex Scout
Browser automation fails or Chrome not found
Cortex Scout auto-detects Chrome, Chromium, or Brave. If detection fails, set the CHROME_EXECUTABLE environment variable to the full path of your browser binary (e.g. /usr/bin/google-chrome).
MCP client shows the server as disconnected or crashing
Check that RUST_LOG is set to 'warn' — the 'info' level floods stderr with output that can confuse MCP clients expecting clean stdio. Also verify the binary path in your config is absolute and the file is executable.
deep_research returns no results or errors
Confirm DEEP_RESEARCH_ENABLED=1 is set and that OPENAI_API_KEY and OPENAI_BASE_URL are correct. Without these, the synthesis step fails silently. You can test with OPENAI_BASE_URL pointing to a local Ollama instance.
Frequently Asked Questions about Cortex Scout
What is Cortex Scout?
Cortex Scout is a Model Context Protocol (MCP) server that unified web extraction and stateful automation engine for ai. replaces heavy testing frameworks with token-optimized browser control, deep research, and hitl. It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Cortex Scout?
Follow the installation instructions on the Cortex Scout GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with Cortex Scout?
Cortex Scout works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Cortex Scout free to use?
Yes, Cortex Scout is open source and available under the MIT license. You can use it freely in both personal and commercial projects.
Cortex Scout Alternatives — Similar Browser Automation Servers
Looking for alternatives to Cortex Scout? Here are other popular browser automation servers you can use with Claude, Cursor, and VS Code.
Chrome DevTools MCP
★ 40.6kAI-powered Chrome automation server with natural language element detection. Control Chrome browser through MCP protocol for testing, debugging, and performance analysis. Features 91% accuracy in element location, works with free AI models, and suppo
UI TARS Desktop
★ 34.9k📇 🏠 - Browser automation capabilities using Puppeteer, both support local and remote browser connection.
Playwright
★ 32.8kA production-ready browser automation server that enables AI assistants to interact with web pages using tools for navigation, element interaction, and data extraction. It features a built-in Inspector UI and robust crash recovery for reliable automa
Page Agent
★ 18.0kJavaScript in-page GUI agent. Control web interfaces with natural language.
Chrome
★ 11.7kAn extension-based MCP server that enables AI assistants to control your browser, leveraging existing sessions and login states for automation and content analysis. It provides over 20 tools for semantic tab search, interactive element manipulation,
LAMDA
★ 7.8kThe most powerful Android RPA agent framework, next generation mobile automation.
Browse More Browser Automation MCP Servers
Explore all browser automation servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up Cortex Scout in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use Cortex Scout?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.