Prompt Enhancer

A Chrome extension that rewrites your draft prompt into a well-structured, high-quality query using a local LM Studio model — before you send it to ChatGPT, Claude, Gemini, or Perplexity.

No data leaves your machine. The extension talks directly to your local LLM.

How it works

1. Type a rough prompt in any supported AI chat interface
2. Press Ctrl+Shift+E (or click the Elite Enhance button that appears above the input)
3. The extension sends your draft to your local LM Studio model with a prompt engineering system prompt
4. The improved version replaces your text — ready to send

The system prompt instructs the model to: analyze intent, fill in missing context, add structure (CO-STAR / CREATE frameworks where appropriate), adjust tone, and remove fluff — without changing what you're asking for.

Supported sites

• chatgpt.com
• claude.ai
• gemini.google.com
• perplexity.ai
• localhost:8888 (local dev / Open WebUI)

Installation

1. Clone or download this repo
2. Open Chrome and go to chrome://extensions/
3. Enable Developer mode (top-right toggle)
4. Click Load unpacked and select the prompt-enhancer folder
5. The extension icon appears in your toolbar

Configuration

Click the extension icon to open settings:

Setting	Description	Default
LM Studio base URL	OpenAI-compatible endpoint	http://127.0.0.1:1234/v1
Model name	Model ID as shown in LM Studio	model-identifier
API key	Dummy value is fine for LM Studio	lm-studio
Mode	enhance-only replaces text; enhance-and-send also submits	enhance-only

Works with any OpenAI-compatible server

While designed for LM Studio, the extension works with any server that exposes an OpenAI-compatible /v1/chat/completions endpoint:

• LM Studio — http://127.0.0.1:1234/v1
• Ollama — http://127.0.0.1:11434/v1
• llamafile — http://127.0.0.1:8080/v1

Keyboard shortcut

Ctrl+Shift+E — enhance the current prompt on any supported page.

The floating Elite Enhance button tracks the active input field and repositions automatically as you scroll.

Privacy

All processing happens locally. No prompt text is sent to any external service. The extension only makes requests to the URL you configure in settings (default: localhost).

License

MIT