The macOS shortcut that turns
rough text into AI-ready prompts.
Select any text, press ⌃⌥Space, and PromptRework adds the role, context, constraints, and output format your AI needs — right inside ChatGPT, Claude, Cursor, or anywhere you type.
“The model isn’t your bottleneck. Your prompt is. I was rewriting lazy prompts twenty times a day in Cursor and Claude, so I made it one shortcut.”
Sharper prompts in every AI tool you use.
PromptRework works anywhere you write. Highlight text, press the hotkey, get a better prompt back in the same field — no switching tabs, no copy-paste.
…and every other app with a text field. PromptRework runs system-wide.
From vague request to
a usable prompt.
make me a habit tracker app
Build a habit-tracker web app where I can add daily habits, check them off, and see my streak…
DoneDifferent tasks need
different prompts.
A coding task, a research question, and an image prompt each need different structure. PromptRework recognizes what you’re doing and builds the right kind of prompt.
Not a rewrite.
A better prompt.
PromptRework turns a vague request into a prompt built for the job — code, research, content, images, video, and more.
summarise the latest on rag for me
Research the current state of Retrieval-Augmented Generation (RAG). Depth: intermediate. Assume the reader is a software engineer or ML practitioner familiar with LLMs but wanting a structured update on where the field stands today. Cover in order: (1) a brief definition and why RAG matters; (2) the most significant recent advances (e.g. advanced chunking strategies, re-ranking, hybrid search, agentic RAG, GraphRAG); (3) key limitations and open problems that practitioners still face; (4) notable tools, frameworks, and papers driving current progress. Focus on the current state; include historical context only where it clarifies a recent shift. Cite sources or named papers/frameworks inline where possible. Distinguish established best practices from emerging or contested approaches. Output: a structured report with a header per section, under 600 words.
Better first answers.
Your rough draft, rewritten where you typed it.
Highlight any text on your Mac, press ⌃⌥Space, and PromptRework replaces it with a detailed prompt — without leaving the app you’re in.
It starts with what you are doing.
In Cursor, it recognizes a coding task. In the browser, it can use the page in front of you. The prompt starts with context instead of asking you to repeat it.
More than a longer prompt.
PromptRework adds the detail, structure, and direction that make an AI answer more useful the first time.
Built for the job, not a generic template.
Writing, code, research, images, design — PromptRework recognizes the task and builds the kind of instruction it needs.
Keep your best prompts.
Save and search rewrites that worked. History is optional and encrypted so only you can read it.
Keep shaping it until it fits.
Use Tab for another angle, more detail, or a different direction— without rebuilding the prompt from scratch.
Use AI every day?
Start with a better prompt.
Start free. No credit card. Upgrade when better prompts are part of your daily workflow.
- ✓Unlimited fast rewrites + 10 premium rewrites/month
- ✓Core rewrite modes
- ✓Works across your favorite AI tools
- ✓Refine Loop for quick improvements
- ✓Community support
- ✓Unlimited premium rewrites for the work you do every day
- ✓Advanced rewrite modes for code, writing, research, design & more
- ✓App-aware context that reads the AI tool and page you're in, from ChatGPT and Claude to Cursor, Copilot, and Lovable
- ✓Refine Loop to push outputs further
- ✓Prompt history with encrypted storage
- ✓Faster workflow across every AI tool
- ✓Priority email support
- ✓Everything in Free
Everything else you
might be wondering.
Yes. PromptRework works in any app with a text field — ChatGPT, Claude, Cursor, Gemini, Perplexity, Copilot, Lovable, VS Code, Notion, even your browser's address bar. It reads what you've selected and rewrites it in place.
Your selected text is sent over HTTPS, rewritten by Anthropic's Claude, and discarded. Nothing is stored unless you turn on History — which is encrypted so that even we can't read it.
No. Rewrites run through Anthropic's API, which does not use API data to train models. We don't train anything on your text. It's processed, returned, and discarded.
Accessibility lets PromptRework read your selected text and write the improved version back into the same field — that's the whole feature. We chose this over the more invasive Input Monitoring permission: PromptRework never sees your keystrokes, only text you explicitly invoke it on.
Unlimited fast rewrites and 10 premium (Sonnet) rewrites per month. Use it in your normal workflow — upgrade to Pro when you want unlimited premium quality.
Yes. macOS 14 (Sonoma) and newer. Universal binary — runs natively on Apple Silicon and Intel. Windows and Linux aren't planned yet.
ChatGPT's helper only works inside ChatGPT. PromptRework works across every AI tool and app you use — one shortcut, no copy-paste, no switching tabs.
Yes. One click in your account cancels at the end of your current billing period. No emails, no questions asked.