A lightweight VS Code extension for optimizing AI prompts using configurable LLM providers.
Features
Model Configuration — Support for OpenAI, Gemini, DeepSeek, and custom providers with secure API key storage.
Optimization Types — Choose between Basic, Professional, and Step-by-Step optimization modes, each with dedicated prompts.
Full Internationalization (i18n) — All UI labels support Chinese/English switching based on language settings.
One-Click Optimization — Paste or type your prompt, click "Optimize", and get an improved version instantly.
Optimize from Clipboard — Copy text anywhere (editor, Webview, external app), press Ctrl+Alt+M (⌘⌥M on Mac), and paste the optimized result.
Editor Integration — Right-click selected text in the editor to send it directly to the optimizer sidebar.
One-Click Copy — Copy the optimized prompt to your clipboard with a single click.
Getting Started
Installation
Open VS Code
Press Ctrl+Shift+X to open the Extensions panel
Search for Prompt Alchemy
Click Install
Configuration
Click the Prompt Alchemy icon in the Activity Bar to open the sidebar
Expand the Model Configuration section
Select your Provider (OpenAI, Gemini, DeepSeek, or Custom)
Enter your API Key
(Optional) Modify the Base URL if using a custom endpoint
Set the Model Name (e.g., gpt-4o, gemini-1.5-pro, deepseek-chat)
Click Save
Usage
Sidebar Workflow
Open the Prompt Alchemy sidebar
Enter your prompt in the Original Prompt text area
Click Optimize
View the improved prompt in the Optimized Prompt section
Click Copy to copy the result to your clipboard
You can also select text in the editor, right-click, and choose Send to Prompt Alchemy to fill the input automatically.
Clipboard Optimization (Works Everywhere)
This feature lets you optimize text without opening the sidebar — works from any context.
Step
Action
Windows / Linux
macOS
1
Copy the text you want to optimize
Ctrl+C
⌘C
2
Press the shortcut to trigger optimization
Ctrl+Alt+M
⌘⌥M
3
Paste the optimized result
Ctrl+V
⌘V
How it works: The extension reads the clipboard content, sends it to your configured LLM for optimization, and writes the result back to the clipboard.
