✨ Prompt Enhancer Local

Turn your rough, lazy prompt comments into highly detailed, professional prompts — powered by a local AI running on your own machine. No cloud, no API keys, 100% private.

---

📖 What Does This Extension Do?

Imagine you're coding and you write a quick comment like:

// make a login page with validation

You select that text, press a keyboard shortcut, and the extension rewrites it into something like:

Create a responsive login page component using React with TypeScript.
Include email and password input fields with the following validation rules:
- Email must be a valid email format
- Password must be at least 8 characters with one uppercase letter and one number
Use functional components. Style with Tailwind CSS. Use pnpm as the package manager.
Display inline error messages below each field. Add a "Remember Me" checkbox and a
"Forgot Password?" link. Handle form submission with proper loading state...

It takes your rough idea and turns it into a detailed, context-aware prompt that any AI coding assistant can understand perfectly.

---

🧰 Prerequisites (What You Need Before Starting)

Before using this extension, you need one thing installed on your computer: a local AI model server. We recommend Ollama — it's free and easy.

Step 1: Install Ollama

1. Go to https://ollama.com
2. Click Download and choose your operating system (Mac, Windows, or Linux)
3. Follow the installer — it's a standard app installation, just click "Next" a few times
4. Once installed, Ollama runs quietly in the background (you'll see a small llama icon in your menu bar on Mac)

Step 2: Download an AI Model

Open your Terminal app (on Mac, search for "Terminal" in Spotlight) and type:

ollama pull llama3

This downloads the llama3 model (~4 GB). Wait for it to finish — it only needs to download once.

💡 Tip: If you have a powerful computer with lots of RAM (32 GB+), you can try larger models like llama3:70b for even better results. If your computer is older, try the smaller llama3:8b.

Step 3: Verify Ollama Is Running

In your Terminal, type:

ollama list

You should see llama3 (or whichever model you downloaded) in the list. If you do, you're all set!

---

📦 Installing the Extension

Step 1: Download the Extension

1. Go to the GitHub Releases page
2. Find the latest release (e.g., v0.1.0)
3. Under Assets, click on prompt-enhancer-local-0.1.0.vsix to download it
4. Save it somewhere easy to find (like your Downloads folder)

Step 2: Install in Your IDE

For VS Code:

1. Open VS Code
2. Open the Command Palette — press Cmd + Shift + P (Mac) or Ctrl + Shift + P (Windows/Linux)
3. Type "Install from VSIX" — you'll see an option called "Extensions: Install from VSIX...", click it
4. A file browser will open — navigate to where you saved the .vsix file and select it
5. Wait a few seconds — you'll see a notification saying the extension was installed
6. Click "Reload Now" (or restart VS Code) to activate the extension

For Antigravity IDE / Cursor / Other VS Code Forks:

The process is the same — open the Command Palette (Cmd + Shift + P), search for "Install from VSIX", and select the downloaded file.

Step 3: Verify Installation

1. Open the Extensions sidebar — click the puzzle piece icon on the left sidebar, or press Cmd + Shift + X (Mac)
2. Search for "Prompt Enhancer"
3. You should see it listed as installed ✅

That's it! No cloning, no terminal commands, no compiling. The extension is now permanently installed and ready to use.

---

How to Use the Extension

The Basic Workflow

1. Open any code file in VS Code / Antigravity IDE

2. Type a rough prompt anywhere — as a comment, inline, wherever you like

   # add error handling to this function with retries

3. Select (highlight) that text with your mouse or keyboard

4. Press the keyboard shortcut:

   Operating System	Shortcut
   Mac	Cmd + Shift + R
   Windows / Linux	Ctrl + Shift + E

5. Wait a moment — you'll see a notification saying "Enhancing prompt…" at the bottom-right

6. Done! — your selected text is replaced with a detailed, professional prompt

What Happens Behind the Scenes

You type:         "add dark mode toggle"
                        ↓
Extension reads:   File language (e.g., TypeScript)
                   Project rules from .enhancerc.json
                        ↓
Sends to Ollama:   A carefully constructed instruction asking the AI
                   to expand your rough text into a detailed prompt
                        ↓
You get back:      A multi-line, detailed prompt that respects your
                   project's specific tools and conventions

---

⚙️ Configuring the Extension

Project-Level Configuration (.enhancerc.json)

You can create a file called .enhancerc.json in the root folder of any project to tell the extension about your project's rules and conventions. The enhanced prompts will then automatically include these rules.

1. In your project's root folder, create a file named .enhancerc.json
2. Add your project-specific rules as key-value pairs:

{
  "packageManager": "pnpm",
  "styling": "Tailwind CSS",
  "framework": "Next.js 14 with App Router",
  "testing": "Vitest + React Testing Library",
  "customRules": "Use functional components. Prefer server components when possible."
}

💡 A sample file is included as .enhancerc.example.json — you can copy and rename it:

cp .enhancerc.example.json .enhancerc.json

You can add any rules that are relevant to your project. Common examples:

Key	Example Value
packageManager	"npm", "pnpm", "yarn"
styling	"Tailwind CSS", "CSS Modules", "styled-components"
framework	"Next.js 14", "React 18", "Vue 3"
language	"TypeScript strict mode", "Python 3.12"
testing	"Jest", "Vitest", "Pytest"
customRules	Any free-text instructions for the AI

If this file doesn't exist, the extension works fine — it just won't include project-specific rules.

Extension Settings (Global)

You can also configure global settings in VS Code:

1. Open Settings — press Cmd + , (Mac) or Ctrl + , (Windows/Linux)
2. Search for "Prompt Enhancer"
3. You'll see two settings:

Setting	Default	What It Does
LLM Endpoint	http://localhost:11434/api/generate	The URL where Ollama (or another local LLM) is running. Change this only if you run Ollama on a different port or machine.
Model	llama3	The AI model to use. Change this if you downloaded a different model (e.g., mistral, codellama, llama3:70b).

---

🔧 Changing the Keyboard Shortcut

If the default shortcut conflicts with another extension or you simply prefer a different one:

Method 1: Via VS Code Settings (Easiest)

1. Open Keyboard Shortcuts — press Cmd + K, Cmd + S (Mac) or Ctrl + K, Ctrl + S
2. Search for "Prompt Enhancer"
3. Click the ✏️ pencil icon next to the command
4. Press your desired key combination
5. Press Enter to save

Method 2: Edit package.json Directly

Open package.json and find the keybindings section:

"keybindings": [
    {
        "command": "promptEnhancer.enhanceSelectedText",
        "key": "ctrl+shift+e",
        "mac": "cmd+shift+r",
        "when": "editorTextFocus"
    }
]

Change the key (Windows/Linux) and mac values to your preferred combination. Examples:

Shortcut	Value to Use
Cmd + Alt + E	"cmd+alt+e"
Cmd + K, then E	"cmd+k cmd+e"
Ctrl + Shift + P	"ctrl+shift+p"

---

❓ Troubleshooting

"Could not connect to local LLM. Is Ollama running?"

This means the extension can't reach Ollama. Try:

1. Check if Ollama is running — look for the llama icon in your menu bar (Mac) or system tray (Windows)
2. Start Ollama manually — open Terminal and run:

   ollama serve

3. Test the connection — run this in Terminal:

   curl http://localhost:11434/api/generate -d '{"model": "llama3", "prompt": "Hi", "stream": false}'

   If you get a JSON response back, Ollama is working. If not, reinstall Ollama.

"Please select some text first."

You pressed the shortcut without highlighting any text. Select (highlight) the text you want to enhance first, then press the shortcut.

The response takes too long

Local LLMs depend on your hardware. If it's slow:

• Try a smaller model: ollama pull llama3:8b
• Update the Model setting in VS Code to match: llama3:8b
• Close other heavy applications to free up RAM

The enhanced prompt doesn't match my project style

Create or update the .enhancerc.json file in your project root with more specific rules. The more detail you provide, the better the output.

---

📁 Project Structure

promptEnhancer/
├── .vscode/
│   ├── launch.json          # Debug launch configuration (F5)
│   └── tasks.json           # Build task for compilation
├── src/
│   ├── extension.ts         # Main extension logic
│   ├── configReader.ts      # Reads .enhancerc.json configuration
│   └── llmService.ts        # Communicates with the local LLM
├── out/                     # Compiled JavaScript (auto-generated)
├── package.json             # Extension manifest & settings
├── tsconfig.json            # TypeScript configuration
├── .enhancerc.example.json  # Sample project rules configuration
└── README.md                # This file

---

�️ For Developers (Contributing / Modifying the Code)

If you want to modify the extension or contribute to it, follow these steps:

1. Clone & Install

git clone https://github.com/dotheki/prompt-enhancer-local.git
cd prompt-enhancer-local
npm install        # or: pnpm install

2. Run in Debug Mode

1. Open the project folder in VS Code / Antigravity IDE
2. Press F5 — this opens a second IDE window (called the Extension Development Host) with your extension loaded
3. Make changes to the code in src/, then press Ctrl+Shift+F5 to restart and test again
4. You can set breakpoints, view console.log output in the Debug Console, and step through the code

3. Repackage After Changes

npm run compile
pnpm dlx @vscode/vsce package --allow-missing-repository

This creates a new .vsix file you can share or install.

---

�📜 License

MIT — use it however you like.