Pick a proven prompt, fill in the blanks, and copy the finished version. Or write your own with {{variables}}.
Variables found in the template appear here as fields.
This is a library of ready-to-use prompt templates for ChatGPT, Claude, Gemini, and any other text model. Pick a template, fill in the blanks, and copy a finished prompt. Every template uses double-brace placeholders like {{topic}}, and the tool turns each one into a field you fill in. You can also write your own template from scratch. Everything runs in your browser, and nothing you type is stored or sent anywhere.
A prompt template is a reusable prompt with the specifics pulled out into variables. Instead of rewriting "summarize this article about climate policy in three bullet points" every time, you keep one template, summarize {{text}} in {{number}} bullet points, and change only the blanks. Templates capture the structure that makes a prompt work, the role, the task, the context, and the output format, so you get consistent results without starting from a blank box each time.
The best prompts share four parts. A role tells the model who to be: "You are an experienced copy editor." A task states exactly what to do: "Rewrite the text below for clarity." Context supplies the material and any constraints: the text itself, the audience, the tone. And a format specifies the shape of the answer: "Return a bulleted list" or "Reply with only the corrected text." The templates here are built on this skeleton, which is why they produce focused answers instead of rambling ones.
Writing a prompt from scratch every time is slow and inconsistent. You forget to specify the format, or the tone drifts, or you leave out the context the model needed. A template locks in the parts that worked and exposes only the parts that change. That makes your results repeatable, which matters most when you run the same kind of task often: weekly summaries, code reviews, outreach emails, study notes. Save the templates you like and you build a personal prompt toolkit.
The built-in templates span the most common uses: writing (summarize, rewrite, outline, proofread), coding (explain code, find the bug, write tests), marketing (headlines, product descriptions, ad copy), email (cold outreach, polite follow-up, reply drafting), and learning (explain a concept simply, quiz me, summarize to study). Each is a starting point you can edit. They are model-agnostic, so the same template works in ChatGPT, Claude, or any other assistant.
Type any prompt into the template box and wrap the parts that change in double braces. "Write a {{tone}} {{platform}} post about {{topic}} in {{number}} words" creates four fields automatically. Use clear, lowercase variable names so the fields read well. Reusing the same variable name fills every copy of it at once. When you are happy with the structure, fill the blanks, copy the finished prompt, and keep the template for next time by saving the text somewhere you can find it.
Three habits raise the quality of almost any prompt. Be specific: "in 100 words, for a beginner" beats "briefly." Show an example: one sample of the output you want teaches the model faster than a paragraph of description. Name the format: ask for a table, a list, or "only the code" so you do not have to clean up the answer. The templates here bake these habits in, but they help even more when you write your own. Counting tokens with a token counter also helps you keep long prompts inside a model's context window.
Choose a template from the menu and its text loads into the editable box. The tool scans for {{placeholders}}, removes duplicates, and builds a labeled field for each one. As you type into the fields, the finished prompt updates live, with any blank you have not filled left visible so you can see what is missing. Copy the result into your AI assistant of choice. Everything happens in your browser, so your prompts and the details you type stay on your device. You can confirm that in your browser network panel.
It is a reusable prompt with the specifics pulled out into variables. You keep the structure that makes the prompt work and change only the blanks, so you get consistent results without rewriting the whole prompt each time.
All of them. The templates are plain text and model-agnostic, so the same one works in ChatGPT, Claude, Gemini, or any other text assistant. Paste the finished prompt wherever you chat with the model.
Anything wrapped in double braces, like {{topic}}, becomes a fillable field. The tool detects them automatically, removes duplicates, and fills every copy of a repeated variable at once as you type. Blanks you leave empty stay visible in the output so you know what is missing.
Yes. Type any prompt into the template box and wrap the parts that change in double braces. The fields appear instantly. Use clear lowercase variable names, and reuse a name to fill it in several places at once.
Be specific about length, audience, and tone; show one example of the output you want; and name the format you expect, such as a list or a table. These three habits improve almost any result, and the built-in templates apply them for you.
No. The tool runs entirely in your browser. The templates and everything you type into the fields stay on your device, with nothing sent to a server, logged, or stored. You can verify by watching your browser network tab stay empty.