Qwen3 MCP

The Qwen3 MCP Server is a powerful local LLM code proxy tool that provides over 80 tools to support comprehensive coding capabilities such as file operations, command execution, Git management, web search, and memory planning. It supports two modes: HTTP browser chat and LM Studio MCP integration.

Developer tools Artificial intelligence chatbots #Code Proxy #Local LLM #Multi-tool Integration #MCP Service .JavaScript

rating : 2 points

downloads : 4.8K

update time : 2026-03-13

Open Site

What is the Qwen3 MCP Server?

The Qwen3 MCP Server is a local AI programming assistant system that transforms the Qwen3 model in LM Studio into a fully functional programming agent. By providing over 80 practical tools, it enables the AI to directly operate your computer files, run commands, manage Git repositories, search the web, create blogs, etc., just like a real programming assistant.

How to use the Qwen3 MCP Server?

You can use it in two ways: 1) Browser chat mode: Open the chat interface in your browser and have a conversation with the AI assistant; 2) MCP direct mode: Integrate directly in LM Studio, and the AI can directly call tools without an additional interface. Both methods offer the same powerful functionality.

Applicable scenarios

Suitable for developers, technical writers, students, and any users who need programming assistance. Whether you're writing code, managing projects, creating documents, learning new technologies, or automating daily tasks, this assistant can provide practical help.

Main Features

Dual-mode operation

Supports two usage methods: browser chat interface and direct integration in LM Studio, meeting the preferences of different users

Over 80 practical tools

Provides a complete set of tools including file operations, command execution, Git management, web search, memory planning, etc.

Intelligent tool aliases

Automatically handles incorrect tool names that the AI model may call, ensuring correct execution of instructions

Skill system

Comes with 16 professional skill packages built-in and can be extended to install more skills, allowing the AI to learn knowledge in specific domains

GitHub blog tool

Specialized tools to help create, manage, and deploy GitHub Pages blogs

ComfyUI integration

32 ComfyUI workflow tools, supporting AI image generation and editing

Advantages

Runs completely locally, protecting privacy and data security

Most functions can be used without an internet connection

Over 80 tools cover all daily development needs

Intelligent alias system reduces AI instruction errors

Scalable skill system continuously enhances AI capabilities

Dual-mode design adapts to different usage habits

Limitations

Requires local installation of Node.js and LM Studio

Reading large files may be limited by WebSocket timeout

Some network functions require an internet connection

Skill installation requires manual operation

The configuration process may be complex for non-technical users

How to Use

Installation Preparation

Ensure that Node.js 18+ and LM Studio are installed, and the Qwen3 model is downloaded

Quick Installation

Run the installation script to automatically configure all necessary components

Select Usage Mode

Choose between browser mode or LM Studio direct mode according to your needs

Start Using

Enter your requirements in the chat interface, and the AI will call the corresponding tools to help you complete the task

Usage Examples

Code Project Creation and Management

The AI helps you create new projects, manage file structures, run tests, and commit code to Git

Technical Blog Writing

Use the GitHub blog tool to create and publish technical articles

Batch File Processing

Perform operations such as batch renaming, search and replace, and file organization

Learning New Skills

Learn knowledge in specific technical fields through the skill system

Frequently Asked Questions

What kind of computer configuration do I need to run this server?

What's the difference between browser mode and MCP mode?

How to install new skills?

What should I do if I encounter a timeout error when reading a large file?

What should I do if the AI calls an incorrect tool name?

How to view all available tools?

Related Resources

LM Studio Official Website

Download LM Studio and AI models

Node.js Download

Download the Node.js runtime environment

GitHub Skill Repository

More installable skill packages

Qwen3 Model Download

Recommended download address for the Qwen3 model

GitHub Pages Documentation

Official documentation for GitHub Pages

ComfyUI Official Website

ComfyUI AI image generation tool

🚀 Qwen3 MCP Server

A complete MCP server that empowers LM Studio's Qwen3 (or any local LLM) with full coding agent capabilities, featuring over 80 tools for file operations, command execution, git, web search, memory, planning, and a comprehensive skills system.

✨ Features

Two Server Modes:
- HTTP Mode - Engage in browser-based chat at http://localhost:3847/chat.html
- MCP Mode - Seamlessly integrate with LM Studio's MCP interface
80+ Tools:
- File operations: read, write, edit, search, glob
- Command execution: shell, background processes
- Git operations: status, diff, commit, branch, push, pull
- Web tools: search, image search, URL fetch
- Memory & planning: notes, scratchpads, task lists, plans
- Media: read images, PDFs, screenshots
- Jupyter notebook support
- ComfyUI integration: 32 workflow tools
- GitHub Blog: Jekyll blog creation, themes, deployment
Tool Name Aliases: Server-side routing effectively handles common model hallucinations (e.g., edit -> edit_file, bash -> execute_command)
Skills System: Comes with 16 pre-installed skills, and you can install more from awesome-agent-skills

🚀 Quick Start

HTTP Mode (Browser Chat)

# Start the server
start-chat.bat

# Open in browser
http://localhost:3847/chat.html

MCP Mode (LM Studio Direct)

Add the following to your LM Studio MCP config (~/.lmstudio/mcp-servers.json):

{
  "mcpServers": {
    "qwen3-mcp": {
      "command": "node",
      "args": ["<YOUR_PATH>/src/index.js"],
      "cwd": "<YOUR_PATH>"
    }
  }
}

Replace <YOUR_PATH> with the actual path where you installed the server. Then restart LM Studio.

📦 Installation

Prerequisites

Node.js 18+
LM Studio with a model loaded (Qwen3 is recommended)

Quick Setup (Windows)

cd <YOUR_PATH>
setup.bat

This setup process will:

Install npm dependencies
Create the required directories
Automatically configure LM Studio MCP

Manual Setup

# Install dependencies
npm install

# Setup MCP config for LM Studio
npm run setup

npm Scripts

Command	Description
`npm run setup`	Configure LM Studio MCP
`npm run start`	Start the HTTP server (port 3847)
`npm run mcp`	Run the MCP server directly (stdio)
`npm run dev`	Start with auto-reload

📚 Documentation

Project Structure

qwen3-mcp/
├── src/
│   ├── index.js           # MCP server (stdio mode)
│   └── tools/             # Tool implementations
│       ├── filesystem.js  # File operations
│       ├── edit.js        # File editing
│       ├── bash.js        # Command execution
│       ├── git.js         # Git operations
│       ├── search.js      # Glob/grep search
│       ├── web.js         # Web search, fetch
│       ├── memory.js      # Notes, scratchpads
│       ├── planning.js    # Plans, task tracking
│       ├── tasks.js       # Todo lists
│       ├── thinking.js    # Reasoning tools
│       ├── context.js     # Conversation context
│       ├── interaction.js # User prompts
│       ├── media.js       # Images, PDFs
│       ├── notebook.js    # Jupyter support
│       ├── comfyui.js     # ComfyUI workflows
│       ├── github-blog.js # Jekyll blog tools
│       └── skills.js      # Skills system
├── frontend/
│   ├── server.js          # HTTP server (port 3847)
│   ├── chat.html          # Browser chat interface
│   └── index.html         # Image viewer interface
├── skills/                # Installed skills (16)
│   ├── chrome-extension/  # Chrome extension development (MV3)
│   ├── code-review/       # Code review methodology
│   ├── comfyui-nodes/     # ComfyUI custom node development
│   ├── comfyui-workflow/  # ComfyUI workflow creation
│   ├── differential-review/ # Security-focused diff review
│   ├── docx/              # Word document creation
│   ├── frontend-design/   # Frontend UI/UX
│   ├── github-blog/       # Jekyll blog for GitHub Pages
│   ├── mcp-builder/       # Build MCP servers
│   ├── modern-python/     # Python tooling (uv, ruff)
│   ├── react-best-practices/ # React patterns (Vercel)
│   ├── shadcn-ui/         # Modern component library
│   ├── static-analysis/   # CodeQL, Semgrep, SARIF
│   ├── testing-handbook-skills/ # Fuzzers, sanitizers
│   ├── web-artifacts-builder/ # HTML/React prototypes
│   └── web-design-guidelines/ # UI/UX fundamentals
├── start-chat.bat         # Start HTTP server
├── stop-chat.bat          # Stop server
├── restart-chat.bat       # Restart server
└── install-skill.bat      # Install skills from GitHub

Available Tools (80+)

File Operations

Tool	Description
`read_file`	Read file contents with line numbers (params: `file_path`, `offset`, `limit`)
`write_file`	Write/create files (params: `file_path`, `content`)
`edit_file`	Find and replace in files (params: `file_path`, `old_string`, `new_string`)
`list_directory`	List directory contents (params: `path`)
`create_directory`	Create directories (params: `path`)
`delete_file`	Delete files (params: `path`)
`move_file`	Move/rename files (params: `source`, `destination`)
`copy_file`	Copy files (params: `source`, `destination`)
`file_info`	Get file metadata (params: `path`)
`get_working_directory`	Get the current working directory
`set_working_directory`	Set the working directory (params: `path`)

Edit Tools

Tool	Description
`insert_at_line`	Insert at a specific line (params: `file_path`, `line`, `content`)
`replace_lines`	Replace a line range (params: `file_path`, `start_line`, `end_line`, `content`)
`append_to_file`	Append to a file (params: `file_path`, `content`)
`prepend_to_file`	Prepend to a file (params: `file_path`, `content`)

Search

Tool	Description
`glob_search`	Find files by pattern (params: `pattern`, `cwd`)
`grep_search`	Search file contents with a regex (params: `pattern`, `path`)
`find_definition`	Find code definitions (params: `name`, `path`)

Command Execution

Tool	Description
`execute_command`	Run shell commands (params: `command`, `cwd`, `timeout`)
`execute_background`	Run commands in the background (params: `command`, `cwd`)
`read_output`	Read background process output (params: `session_id`)
`kill_session`	Kill a background process (params: `session_id`)
`list_sessions`	List running processes

Git

Tool	Description
`git_status`	Get the repository status
`git_diff`	Show changes
`git_log`	View the commit history
`git_add`	Stage files
`git_commit`	Create commits
`git_branch`	List/create branches
`git_checkout`	Switch branches
`git_push`	Push to a remote repository
`git_pull`	Pull from a remote repository
`git_clone`	Clone repositories

Web

Tool	Description
`web_search`	Perform a DuckDuckGo search (params: `query`)
`web_image_search`	Conduct a Bing image search with download (params: `query`)
`web_fetch`	Fetch webpage content (params: `url`)
`web_fetch_image`	Download an image from a URL (params: `url`)

Memory & Planning

Tool	Description
`memory_store`	Store notes with tags
`memory_recall`	Search notes
`memory_list`	List all notes
`scratchpad_write`	Write to a scratchpad
`scratchpad_read`	Read from a scratchpad
`plan_create`	Create execution plans (params: `goal`, `steps`)
`plan_status`	Check the plan progress
`task_add`	Add todo items
`task_list`	List todos

GitHub Blog

Tool	Description
`blog_init`	Initialize a Jekyll blog for GitHub Pages
`blog_post_create`	Create a new blog post
`blog_page_create`	Create a static page
`blog_category_create`	Create a category page
`blog_post_list`	List all blog posts
`blog_nav_update`	Update the navigation menu
`blog_deploy`	Deploy to GitHub Pages
`blog_config`	Update the blog configuration
`blog_theme`	Apply a theme preset or custom colors
`blog_theme_list`	List available themes
`blog_jekyll_theme`	Apply a Jekyll remote theme

Skills

Tool	Description
`list_skills`	List installed skills
`load_skill`	Load skill instructions
`install_skill`	Install from GitHub

Utilities

Tool	Description
`get_current_time`	Get the current date/time
`calculator`	Evaluate math expressions
`think`	Record thinking notes
`ask_user`	Prompt the user for input

Skills System

Skills are instruction packages that teach the AI specialized tasks. The model automatically detects which skill to load based on your request.

Installed Skills (16)

Code Quality & Security:

code-review - Code review methodology
differential-review - Security-focused diff review
static-analysis - CodeQL, Semgrep, SARIF
testing-handbook-skills - Fuzzers, sanitizers, coverage

Web Development:

react-best-practices - React patterns (Vercel)
web-design-guidelines - UI/UX fundamentals
shadcn-ui - Modern component library
frontend-design - Frontend UI/UX
web-artifacts-builder - HTML/React prototypes

ComfyUI & Creative:

comfyui-nodes - Custom node development (V1 + V3 API)
comfyui-workflow - Workflow creation (SD1.5/SDXL/SD3.5/Flux)

Development Tools:

chrome-extension - Chrome extension development (MV3)
mcp-builder - Build MCP servers
modern-python - Python tooling (uv, ruff, pytest)
docx - Word document creation
github-blog - Jekyll blogs for GitHub Pages

Install More Skills

install-skill.bat https://github.com/anthropics/skills/tree/main/skills/pptx

Using Skills

Skills automatically load when they match your request. You can also load them manually:

"What skills do I have?"
"Load the comfyui-nodes skill"
"Use the docx skill to create a report"

Configuration

HTTP Server Port

Edit frontend/server.js:

const PORT = 3847;

Environment Variables

Variable	Default	Description
`IMAGE_DOWNLOAD_DIR`	`~/lmstudio-images`	Image download location

LM Studio Setup (HTTP Mode)

Load a model (Qwen3 is recommended)
Enable the API server (default: localhost:1234)
Open http://localhost:3847/chat.html
Enter the API token if authentication is enabled

LM Studio Setup (MCP Mode)

Add the MCP config (see Quick Start above)
Restart LM Studio
Tools will appear automatically in the model context

API Endpoints (HTTP Mode)

Endpoint	Method	Description
`/`	GET	Image viewer UI
`/chat.html`	GET	Chat interface
`/tool`	POST	Execute a tool `{name, args}`
`/skills`	GET	List skills (JSON)
`/skill?name=X`	GET	Get skill details

Testing

# Test a tool
curl -X POST http://localhost:3847/tool \
  -H "Content-Type: application/json" \
  -d '{"name":"get_current_time","args":{}}'

# List skills
curl http://localhost:3847/skills

🔧 Troubleshooting

Server Issues

# Check if running
curl http://localhost:3847/skills

# Restart
restart-chat.bat

# Force stop
stop-chat.bat

MCP Connection Issues

Check the LM Studio console for errors
Verify that the path in mcp-servers.json is correct
Run node src/index.js manually to test

read_file Timeouts / WebSocket Errors

The read_file tool defaults to 500 lines to prevent LM Studio WebSocket timeouts. For large files, use pagination:

read_file with offset=1, limit=100    # Lines 1-100
read_file with offset=101, limit=100  # Lines 101-200

Tool Name Errors

If the model calls incorrect tool names (e.g., edit instead of edit_file), the server has built-in aliases to handle common mistakes. If you still encounter errors, update the system prompt in LM Studio — refer to SYSTEM_PROMPT.md for the correct prompt.