MCP Bun Server
Provides AI assistants with powerful tools to execute, optimize, and manage JavaScript/TypeScript projects using the Bun runtime.
README
MCP Bun Server
A comprehensive Model Context Protocol (MCP) server implementation optimized for the Bun JavaScript runtime. This server provides AI assistants with powerful tools to execute, optimize, and manage JavaScript/TypeScript projects using Bun's high-performance runtime.
🔄 Runtime Compatibility: The server requires bun to be installed in the environment the MCP server will run (Windows native, Windows WSL2, etc). See more installation details below.

Features
🚀 Bun-Optimized Execution
- Fast Script Execution: Run JavaScript/TypeScript files with Bun's optimized runtime
- Built-in TypeScript Support: Execute TypeScript directly without compilation
- Memory Optimization: Use
--smolflag for memory-constrained environments - Hot Reloading: Development server with
--hotflag support
🔧 Development Tools
- Package Management: Install dependencies with
bun install - Script Runner: Execute package.json scripts with
bun run - Build System: Optimize projects with
bun buildincluding minification and bundling - Test Runner: Fast testing with
bun testand coverage reporting
📊 Performance Analytics
- Project Analysis: Analyze bundle sizes, dependencies, and runtime performance
- Benchmarking: Compare script performance with different optimization flags
- Optimization Suggestions: Get recommendations for Bun-specific optimizations
🖥️ Server Management
- Background Servers: Start and manage long-running Bun/Node.js servers
- Process Monitoring: Track server status, logs, and performance
- Hot Reloading Servers: Development servers with file watching capabilities
🔍 Resource Discovery
- Script Listing: Browse available npm/package.json scripts
- Project Structure: Understand project dependencies and configuration
Quick Start
Prerequisites
- Bun v1.0.0 or later (recommended)
- Node.js v18.0.0 or later (for compatibility)
For MacOS/Linux users, install Bun using the instructions on the Bun website.
For Windows, the installation depends if you develope with WSL2 or not:
- For native Windows development, install Bun using the Windows installer.
- If you develop with WSL2, install Bun in your WSL2 environment using the Linux installation instructions.
Configuration
On VSCode MCP Client
For quick installation, use one of the one-click install buttons below...
Or use the configs below.
If using globally, add the following to your MCP client configuration on settings.json:
{
...
"mcp": {
"servers": {
"bun": {
"command": "bunx",
"args": ["-y", "mcp-bun@latest"],
"env": {
"DISABLE_NOTIFICATIONS": "true"
}
}
}
}
}
On Windows, this doesn't work properly when using WSL2 if your project lives in the WSL2 filesystem and you run VSCode natively on Windows. This happens because the MCP server runs on Windows and the Bun commands are executed in the WSL2 environment, which can lead to path issues.
In this case, configure the MCP server in the project workspace configuration file which makes the MCP server run in the WSL2 environment and execute the Bun commands there.
Create a file named .vscode/mcp.json in your project root with the following content:
{
"servers": {
"bun": {
"command": "bunx",
"args": ["-y", "mcp-bun@latest"],
"env": {
"DISABLE_NOTIFICATIONS": "true"
}
}
On Claude Desktop
Configure your Claude Desktop MCP client with the following:
{
"mcpServers": {
"bun": {
"command": "bunx",
"args": ["-y", "mcp-bun@latest"],
"env": {
"DISABLE_NOTIFICATIONS": "true", // Optional: disable permission prompts
}
}
}
}
For Development/Local Installation
Clone the repository to your local machine, install dependencies, and build the project:
git clone https://github.com/carlosedp/mcp-bun.git
cd mcp-bun
bun install
bun run build
Then configure your MCP client to use the local build:
{
"servers": {
"bun-dev": {
"command": "bun",
"args": ["/home/user/mcp-bun/dist/mcp-bun.js"],
"env": {
"DISABLE_NOTIFICATIONS": "true"
}
},
}
}
For testing there's also the MCP Inspector available, which allows you to run the server with Bun and inspect the commands being executed:
bun run dev
Available Tools
Core Execution Tools
run-bun-script-file
Execute JavaScript/TypeScript files with Bun runtime optimizations.
Parameters:
scriptPath: Path to the script filebunArgs: Optional Bun flags (e.g.,--smol,--hot)args: Arguments to pass to the scriptstdin: Optional standard inputcwd: Working directorytimeout: Execution timeout
run-bun-eval
Execute JavaScript/TypeScript code directly with Bun eval.
Parameters:
code: Code to executeevalDirectory: Execution directorybunArgs: Bun optimization flagsstdin: Standard inputtimeout: Execution timeout
Package Management
run-bun-install
Install dependencies using Bun's fast package manager.
Parameters:
packageDir: Directory containing package.jsondependency: Specific package to install (optional)
run-bun-script
Execute npm scripts using Bun.
Parameters:
packageDir: Directory containing package.jsonscriptName: Script name to runargs: Additional arguments
Build & Optimization
run-bun-build
Build and optimize projects with Bun's bundler.
Parameters:
entryPoint: Entry file to buildoutDir: Output directorytarget: Build target (browser,bun,node)minify: Enable minificationsourcemap: Generate source mapssplitting: Enable code splitting
run-bun-test
Execute tests with Bun's fast test runner.
Parameters:
testPath: Test file or directorycoverage: Enable code coveragewatch: Enable watch modebail: Stop after N failurestimeout: Test timeout
Performance Analysis
analyze-bun-performance
Comprehensive project performance analysis.
Parameters:
projectDir: Project directoryentryPoint: Entry point to analyzeoptions: Analysis options (bundle, dependencies, runtime)
benchmark-bun-script
Benchmark script performance with different optimization flags.
Parameters:
scriptPath: Script to benchmarkiterations: Number of test runswarmup: Warmup runs
Server Management
start-bun-server
Start optimized Bun servers with hot reloading and watch capabilities.
Parameters:
scriptPath: Server script pathcwd: Working directorybunArgs: Bun flagsoptimizations: Hot reload, watch, smol mode options
Additional Server Tools
start-node-server: Start Node.js servers for compatibilitylist-servers: List all running servers with status and logsstop-server: Stop running servers gracefully or forcefullyget-server-logs: Retrieve server logs with filtering options
Version Management
get-bun-version
Get current Bun version and revision information.
list-bun-versions
List available Bun installations.
select-bun-version
Select specific Bun version for execution.
Performance Optimization Tips
Memory Optimization
Use the --smol flag for memory-constrained environments:
bun --smol your-script.js
Security Considerations
- The server prompts for permission before executing any command
- Scripts run with the same permissions as the MCP server process
- Use environment variable
DISABLE_NOTIFICATIONS=truefor automation - Be cautious when running scripts from untrusted sources
Development
Building
bun run build # Build with Bun
bun run build:node # Build with Node.js/TypeScript
Development Mode
bun run dev # Run with Bun + MCP Inspector
bun run dev:node # Run with Node.js + MCP Inspector
Linting
bun run lint # Check for issues
bun run lint:fix # Fix auto-fixable issues
Contributing
- Fork the repository
- Create a feature branch
- Make your changes
- Add tests if applicable
- Submit a pull request
License
This project is licensed under the MIT License - see the LICENSE file for details.
Acknowledgments
- Built on the Model Context Protocol
- Optimized for Bun runtime
- Heavily inspired by the MCP Node.js server
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。