Kaboom Browser AI Devtools MCP
MCP server for browser debugging, inspection, and verification that streams console logs, network errors, and user actions into AI coding assistants.
README
Branch Policy (Read First)
If you want something working, load code and run the server from
STABLE.UNSTABLEmakes zero promises on regressions or issues and is treated as work in progress. Stable builds are compressed, tagged, and moved toSTABLE.
<div align="center">
<img src="docs/assets/images/chrome_store/readme-banner-kaboom.svg" alt="KaBOOM! โ Browser debugging, inspection, and verification for AI coding assistants via MCP" width="100%" />
Kaboom is an AI debugger, inspector, and verification toolkit for local-first browser development workflows. Stream console logs, network failures, exceptions, recordings, and browser evidence into any MCP-compatible coding assistant.
Documentation โข Quick Start โข Features โข MCP Setup
</div>
<div align="center">
๐ฆ Latest Release
Current version: v0.8.2 โ Structured telemetry, session analytics, KaBOOM! branding, and contract-compliant metrics reporting.
macOS / Linux:
curl -sSL https://raw.githubusercontent.com/brennhill/Kaboom-Browser-AI-Devtools-MCP/STABLE/scripts/install.sh | bash
Windows (PowerShell):
irm https://raw.githubusercontent.com/brennhill/Kaboom-Browser-AI-Devtools-MCP/STABLE/scripts/install.ps1 | iex
</div>
Quick Start
Fire up Kaboom (binary + extension + auto-config) in one command:
macOS / Linux:
curl -sSL https://raw.githubusercontent.com/brennhill/Kaboom-Browser-AI-Devtools-MCP/STABLE/scripts/install.sh | bash
Windows (PowerShell):
irm https://raw.githubusercontent.com/brennhill/Kaboom-Browser-AI-Devtools-MCP/STABLE/scripts/install.ps1 | iex
This script automatically:
- Downloads the latest stable binary for your platform.
- Installs the browser extension files to
~/.kaboom/extension. - Auto-configures all detected MCP clients (Claude Code, Cursor, Windsurf, Zed, etc.).
Step 1: Finalize Browser Extension
- Open
chrome://extensions - Enable Developer mode (top right)
- Click Load unpacked
- Select the folder:
~/.kaboom/extension(or wherever the script printed)
Step 2: Restart Your AI Tool
Restart Claude Code, Cursor, Windsurf, or Zed. The Kaboom server will now start automatically when needed.
Full setup guide โ | Per-tool install guide โ
Why Teams Use Kaboom
No debug port required. Other tools need Chrome launched with --remote-debugging-port, which disables security sandboxing and breaks your normal browser workflow. Kaboom uses a standard extension, so your browser stays secure and unmodified.
Single binary, zero runtime. One Go binary that runs anywhere โ no runtime dependencies, no Puppeteer, no framework.
Captures what others can't. WebSocket messages, full request/response bodies, user action recording, Web Vitals, automatic regression detection, visual annotations, and Playwright test generation from real browser sessions โ features no other MCP browser tool offers.
Works with every MCP tool. Claude Code, Cursor, Windsurf, Zed, Claude Desktop, VS Code + Continue. Switch AI tools without changing your debugging setup.
Enterprise-safe by design. Binds to 127.0.0.1 only. Auth headers are stripped automatically. No accounts, no cloud. Anonymous usage stats only (see Privacy). Audit the source โ it's AGPL-3.0.
What It Does
- Console logs โ
console.log(),.warn(),.error()with full arguments - Network errors โ Failed API calls (4xx, 5xx) with response bodies
- Exceptions โ Uncaught errors with full stack traces
- WebSocket events โ Connection lifecycle and message payloads
- Network bodies โ Request/response payloads for API debugging
- User actions โ Click, type, navigate, scroll recording with smart selectors
- Web Vitals โ LCP, CLS, INP, FCP with regression detection
- DOM inspection โ Query the page with CSS selectors via MCP
- Accessibility audits โ WCAG checks with SARIF export
- Security audits โ Credentials, PII, headers, cookies, third-party analysis
- Browser automation โ Click, type, select, upload, navigate with semantic selectors
- Visual annotations โ Draw mode for user feedback with computed style extraction
- Test generation โ Playwright tests from context, self-healing selectors, failure classification
- Reproduction scripts โ Playwright scripts from recorded user actions
- Noise filtering โ Auto-detect and dismiss irrelevant errors
- Developer API โ
window.__kaboom.annotate()for custom context
Privacy
All captured data (logs, network, actions) stays 100% local โ nothing leaves your machine. No cloud, no accounts.
We collect anonymous usage statistics (tool call frequency, session duration, error rates) using a random install identifier not linked to your identity. No URLs, prompts, file contents, browsing data, or personal information is collected. Disable with KABOOM_TELEMETRY=off.
Performance
See latest benchmarks for current performance data.
Last benchmarked: 2026-02-09 on darwin/arm64
Known Issues
See docs/core/known-issues.md for current known issues.
Development
make test # Go server tests
node --test tests/extension/*.test.js # Extension tests
make dev # Build for current platform
Release process & quality gates โ ยท Changelog โ
License
AGPL-3.0 โ Free and open source for all use cases.
Artwork, logos, and the Sparky mascot are Copyright (c) Brenn Hill and are not covered by the AGPL. See LICENSE-ARTWORK for details.
<div align="center">
<img src="docs/assets/images/sparky-wave.png" alt="Sparky the Salamander" width="120" />
Fueling rapid development with AI
If you find Kaboom useful, please consider giving it a star.
</div>
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.