Releases: sdirishguy/MCPHelperSublimePlugin
MCPHelper v1.0.1 - Enhanced Features and Documentation
MCPHelper v1.0.1 - Second Release
🎉 Second Release of MCPHelper
A Sublime Text plugin that connects your editor to a local Model Context Protocol (MCP) server, letting you use OpenAI or Gemini-powered tools directly on selected code.
✨ Features
- Generate code from plain-text prompts
- Review code for bugs, security flaws, or improvements
- Refactor code to enhance readability, maintainability, and performance
- Translate code between languages like Python, JavaScript, Bash, etc.
- Configurable settings for MCP server URL, timeouts, and model parameters
- Menu integration for easy access to all commands
- Robust error handling with user-friendly error messages
- Debug mode for troubleshooting
🚀 Installation
Quick Start
- Download the
MCPHelper-v1.0.1.zip
file - Extract to your Sublime Text User packages folder:
- Windows:
%APPDATA%\Sublime Text\Packages\User\
- macOS:
~/Library/Application Support/Sublime Text/Packages/User/
- Linux:
~/.config/sublime-text/Packages/User/
- Windows:
- Restart Sublime Text
- Configure your MCP server (see documentation)
Alternative: Direct GitHub Installation
- Open Sublime Text
- Press
Ctrl+Shift+P
- Type "Package Control: Add Repository"
- Enter:
https://github.com/your-username/MCPHelperSublimePlugin
- Then install via "Package Control: Install Package"
📋 Requirements
- Sublime Text 4 (Build 4000 or higher)
- Python 3.6+ (usually included with Sublime Text)
- A running MCP Server with tools like
llm_generate_code_openai
configured
📖 Usage
-
Select code in any file
-
Press
Ctrl+Shift+P
and choose:MCP: Generate Code
MCP: Review Code
MCP: Refactor Code
MCP: Translate Code
Or use the MCP Helper menu for easy access.
⚙️ Configuration
Open Sublime Text settings:
- Go to
Preferences > Package Settings > MCPHelper > Settings
- Configure your MCP server URL and model parameters
🔧 Troubleshooting
- Ensure your MCP server is running at the configured URL
- Check the console for debug output
- Use the included
test_mcp_connection.py
script to verify connectivity
📄 Documentation
- README.md - Main documentation
- INSTALLATION.md - Detailed installation guide
- PACKAGE_CONTROL.md - Distribution guide
🎯 What's New
- Enhanced error handling and user feedback
- Improved settings integration and menu system
- Better documentation and troubleshooting guides
- Fixed markdown linting issues across all files
- Updated version management and distribution tools
Note: This plugin requires a running MCP Server on your local machine. You must configure your own OpenAI or Gemini API keys in the MCP server.
v1.0.0
This is the first official release of MCPHelper, a lightweight Sublime Text plugin that connects your editor to a local Model Context Protocol (MCP) server. It brings the power of modern LLMs like OpenAI’s GPT-4o and Google Gemini directly into your coding environment — no browser tabs, no distractions.
✨ Features
🔧 Command Palette Integration
Easily run MCP tools like “Generate Code”, “Review Code”, “Refactor Code”, or “Translate Code” on selected text.
⚙️ LLM Tool Support
Uses your own OpenAI or Gemini API keys via your local MCP server.
🧠 Code Intelligence via JSON-RPC
Designed to be flexible and fast — your local server handles authentication, execution, and response parsing.
💬 Fully threaded
All tool calls run in the background to keep the Sublime UI smooth and responsive.
🖼️ Screenshot included
Preview what the plugin looks like in action.
📦 Requirements
This plugin does not work out-of-the-box with OpenAI or Gemini.
To use it, you’ll need:
A local MCP server running on http://localhost:8000/mcp.json/
One or more tools registered (e.g., llm_generate_code_openai)
Your own API keys for OpenAI or Google Gemini set up in the server config
🛠 Example Use Cases
Generate boilerplate code or templates
Review existing code for improvements or security issues
Refactor legacy code for readability and maintainability
Translate code snippets between languages (e.g., Python → Bash)
🗺 Roadmap Highlights (Future Ideas)
💡 Optional tool autodiscovery
📁 Remote/Cloud MCP compatibility
🧪 Custom user prompts and presets
🔄 VS Code and NeoVim ports
Feel free to fork it, open issues, or contribute. Built to scratch my own itch — now shared for the community. 🛠
—
David Donohue
github.com/sdirishguy