Each entry explains what the prompt produces, lists the inputs it expects, and can be copied in one click. Filter by category, difficulty or assistant.
Quick ways into the library: the newest additions, the most copied, or the entries pitched at your level.
The twelve largest subject areas, by number of prompts. Every category has its own page.
Most prompts are plain instructions with no model-specific syntax, so the major assistants each cover nearly the whole library. The counts show where that is not true.
Every prompt names the roles it was written for, and most name several. These are the forty most common; the filter panel reaches the rest.
This prompt is designed to create a comprehensive security policy tailored to a specific subject or entity. It outlines the necessary guidelines, procedures, and protocols to ensure the…
This prompt guides a seasoned programmer to create efficient and well-structured code in a specified programming language to perform a defined action. It emphasizes best practices, thorough testing,…
This prompt guides a seasoned programmer to analyze a specific piece of code that is causing an error. The programmer is tasked with diagnosing the error's root cause,…
This prompt guides a seasoned programmer to conduct a thorough code review of a specified piece of code. The review focuses on evaluating efficiency, readability, maintainability, and adherence…
This prompt guides a seasoned programmer to refactor a specified piece of code, focusing on enhancing its efficiency, readability, and maintainability while preserving its functionality. The programmer is…
This prompt guides the user to create comprehensive documentation for a specific piece of code. It emphasizes clarity and accessibility for both novice and experienced developers, requiring an…
This prompt guides a seasoned programmer to create comprehensive tests for a specific piece of code using a designated testing framework. It emphasizes the importance of understanding the…
This prompt guides the user to compile a detailed list of current cyber threats targeting a specific sector. It emphasizes the need for comprehensive information, including threat types,…
This prompt guides the creation of an interactive escape room experience that emphasizes collaboration between the guide and the player. It focuses on developing a compelling narrative, designing…
This prompt guides a performance review expert to conduct a benchmarking analysis of employee feedback across different departments. It focuses on collecting, analyzing, and comparing feedback data to…
Page 63 of 74
The TiorAI AI Prompt Library is a free collection of 4,043 ready-to-use AI prompts you can copy, paste and adapt. Every entry works with the assistants people actually use — ChatGPT, Claude, Gemini, Microsoft Copilot, Perplexity and the image models — because the prompts are written in plain language rather than in model-specific syntax. There is no prompt pack to download: the full text of each prompt sits on its own page with a one-click copy button beside it.
Four ways in, and they stack. Browse by category to move through subjects like marketing, writing, coding, SEO, design, business and education. Browse by audience to see the prompts written for your role, from marketers and developers to teachers, founders, students and researchers. Browse by AI model to stay with the assistant you already pay for. Then narrow by difficulty, prompt type, or a plain keyword search. Every combination of filters has its own URL, so a view worth keeping can be bookmarked or sent to a colleague.
Each entry carries two pieces of text that do different jobs. The description explains, in ordinary English, what the prompt produces and what you have to supply. The prompt itself is the instruction you paste in, and it is written with bracketed placeholders — [PRODUCT NAME], [TARGET AUDIENCE], [TONE] — marking the inputs to swap for your own details. That structure is what separates a usable prompt template from a screenshot of somebody else's chat: you can see the moving parts before you run it, and change them.
You do not need to study prompt engineering to get something out of this library, although reading a few entries in one category teaches the shape of a good instruction faster than most courses do. Beginner prompts run as written and need one or two inputs. Intermediate prompts, which are the bulk of the collection, expect several inputs and a decision about tone or format. Advanced prompts assume you will rework the structure around your own workflow rather than run them as they stand.
New prompts are added as they are written and checked, across writing, marketing, SEO, sales, customer support, data analysis, image generation, coding and study. Entries that stop working well as the models change are revised rather than left in place — sorting by Recently Updated shows what has been touched most recently. If a prompt returns something generic, the usual cause is a placeholder left unfilled, and the questions beside this text cover the rest.
Each entry has two pieces of text. The description, shown on the card and at the top of the page, explains in plain language what the prompt produces and what you have to supply. The prompt itself is the instruction you paste into an assistant, and it is what the Copy button puts on your clipboard.
They are the inputs you replace. A prompt written as "Based on the provided [RECIPIENT NAME], create…" expects you to swap that placeholder for your own value before sending it. Leaving a placeholder in is the most common reason a prompt returns something generic.
Beginner prompts work as written and need one or two inputs. Intermediate prompts, which are most of the library, expect several inputs and some judgement about tone or format. Advanced prompts assume you will adapt the structure to your own workflow rather than run them as-is.
Almost all of them are plain-language instructions with no model-specific syntax, so they work in ChatGPT, Claude, Gemini and anything comparable. Each entry lists the assistants it was checked against on its own page.
Yes. Category, audience, model, difficulty and type stack together, and every combination has its own URL, so a filtered view can be bookmarked or shared. Active filters appear above the results and can be removed one at a time.
Newest first, so the front of the library reflects what was added most recently. Switch the sort to A-Z to browse by name, or Recently Updated to see entries that have been revised.
