linshenkx/prompt-optimizer

Prompt Optimizer is a tool for rewriting AI prompts so they produce better results. Instead of staring at a one-line instruction and guessing why the model ignored half of it, you paste your prompt in, click optimize, and the tool runs multi-round iterative rewrites that try to make the instruction clearer and more directive. It can optimize both the system prompt that sets up the AI's role and the user prompt that asks a specific question. The project supports several ways to use it: a web app, a desktop app, a Chrome extension, and Docker deployment. You can bring prompts in from scratch, from templates, from local files, or from a separate library called Prompt Garden, and save the ones that work as reusable favorites with version history, examples, and a record of where they came from. An evaluation pipeline lets you analyze a single output or compare results side by side so you can tell whether a rewrite actually improved things, and an advanced testing mode adds context variables, multi-turn conversation tests, and Function Calling integration. An image generation mode also handles text-to-image, image-to-image, and multi-image prompts. You would use it when you spend time hand-tuning prompts and want a structured environment to draft, test, and refine them. It plugs into mainstream AI models including OpenAI, Gemini, DeepSeek, Zhipu AI, SiliconFlow, and MiniMax. Built in TypeScript, processing happens in your own browser and talks to the AI provider directly without going through an intermediate server, and it supports the Model Context Protocol (MCP) for integration with apps like Claude Desktop.