Enterprise DNA
M MCP Servers Developer low

feedthrough/feedthrough

by Various

A debug bridge embedded inside your web app, exposing runtime internals to AI agents via MCP.

F

MCP

feedthrough/feedthrough

Added 15 June 2026

#agentic-web #browser-automation #claude-code #cursor-ide #cypress #devtools #dom-inspector #mcp

Overview

Feedthrough is a TypeScript debug bridge that embeds into a web application and exposes runtime internals to AI agents through the Model Context Protocol (MCP). It acts as a communication layer between live application state and external agentic tools.

Best for

Best for
Developers building AI agents that need live introspection of web application internals

Use cases

  • Inspecting live application state during AI-assisted debugging
  • Feeding real-time runtime data to agentic coding assistants
  • Building custom AI agent integrations that need access to web app internals

How to use

Install

npx @feedthrough/mcp

Tools exposed

  • get_instructions()
  • query_dom(selector)
  • inspect_element(selector, properties?)
  • get_html(selector)
  • get_console_logs(limit?, levels?, match?, since?)
  • get_network_requests(filter?, since?)
  • get_page_info()
  • connection_status()
  • click(selector)
  • fill(selector, value)
  • hover(selector)
  • press_key(selector, key)
  • set_style(selector, properties)
  • set_attribute(selector, name, value)
  • set_text(selector, text)
  • reset_overrides()

Tested with

Claude Code, Cursor

Example client config

{\n  "mcpServers": {\n    "feedthrough": {\n      "command": "npx",\n      "args": ["@feedthrough/mcp"]\n    }\n  }\n}

Notes

Feedthrough is a TypeScript debug bridge that embeds into a web application and exposes runtime internals to AI agents through the Model Context Protocol (MCP). It acts as a communication layer between live application state and external agentic tools.

2 stars on GitHub. Last updated 2026-06-10. Licensed MIT.

Use cases

  • Inspecting live application state during AI-assisted debugging
  • Feeding real-time runtime data to agentic coding assistants
  • Building custom AI agent integrations that need access to web app internals

Pros

  • Direct runtime access enables deep debugging without static analysis
  • Leverages MCP standard for interoperability with multiple agent frameworks
  • Minimal overhead as a lightweight TypeScript module

Cons

  • Very early project (2 stars) with limited community validation
  • Embedding a debug bridge in production apps may pose security risks
  • Requires custom integration effort beyond simple drop-in usage

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.

Pros

  • Direct runtime access enables deep debugging without static analysis
  • Leverages MCP standard for interoperability with multiple agent frameworks
  • Minimal overhead as a lightweight TypeScript module

Cons

  • Very early project (2 stars) with limited community validation
  • Embedding a debug bridge in production apps may pose security risks
  • Requires custom integration effort beyond simple drop-in usage
Free 27-page guide

Get the free Developer’s Field Guide

A 27-page field guide to the AI coding workflow with Claude. Claude Code, MCP servers, the prompt patterns that work, and what to delegate. Free.

Enter your work email. We send it straight over, plus a few short notes worth knowing. Unsubscribe any time.

No spam. Unsubscribe any time.

Running a business, not writing the code? See the MCP servers picked for operators, and get your first one wired up with us.

Operator picks