A prompt engineering tool for all professionals.

When you type a quick, messy thought into an AI chat, the model has to guess a lot, like your role, its own role, your task, your constraints, information that it has and the information that it needs to look for or assume and the output format you need. Those guesses are exactly where things go wrong: made-up details stated with total confidence, two-page essays when you only needed three bullet points, or frustrating generic corporate AI fluff.

The standard industry advice is to “learn prompt engineering” - add system directives, specify XML delimiters, and write elaborate guardrails. But if you work in marketing, HR, sales, legal, finance, or run a business, wrestling with prompt syntax is an unwanted task, not your job. AI was supposed to take care of all this.

Patkan is a lightweight desktop browser extension that does the prompt engineering for you. It activates only when you type // at the end of your sentence inside your AI chat box like ChatGPT, Claude, Gemini etc.

In the blink of an eye, your rough thought is replaced by a structured, high-quality prompt complete with a clear role, context, task boundaries, and output format—tailored specifically to the AI model you are using.

There is no tab switching, no copying and pasting from external prompt libraries, and no complex configuration. Patkan works natively inside ChatGPT, Claude, Google Gemini, Microsoft Copilot, Perplexity, Google AI Studio and more AI assistant models.

The name originates from the Marathi word “पट्कन” (patkan), which translates to “instantly” or “in a snap.”

The tool was born out of personal frustration with the gap between having a quick thought and actually getting a usable, hallucination-free answer from modern AI chatbots. The solution was to type 300 words well structured prompts all the time, which is time consuming and everyone doesn’t know that. Patkan was built to make AI feel ‘intelligent’ again.

Trust is critical when it comes to your browser. The core code behind Patkan is free and open-source under the GNU AGPL-3.0 license on GitHub. Anyone can inspect the code, review the extension manifests, or run their own local copy.

Patkan is built with zero advertising trackers and no user profiling. Your prompts (raw thoughts) are processed solely to return a structured prompt, and an anonymous cache ensures that repeat transforms happen instantly. The extension is completely deaf until you type // and blind to every website not explicitly listed in its permissions.

Patkan is currently completely free to use. To protect server capacity and prevent automated abuse, simple fair-use limits apply:

• Unauthenticated Guest users: Up to 10 prompt transforms per day.
• Signed-in users: Up to 50 prompt transforms per day.