12 everyday prompts you can copy
Twelve copy-and-paste prompts for daily tasks, with the reason each one works, and a short method to write your own: role, context, goal, format, examples.
A good prompt saves time. A vague one wastes it. These twelve prompts cover common tasks. Copy one, replace the parts in [square brackets], and send it. Each has a short note on why it works, so you can adapt it.
They follow the advice in the vendors' own prompting guides: be clear and direct, give context, say what the output should look like, and show an example when it helps. Check the sources below. Models differ, so treat these as starting points and tune the wording.
Warning: Do not paste secrets, client data or private details. See what not to paste. Models can also state false facts with confidence (hallucination), so check anything that matters.
The 12 prompts
Shorten an email
Use it for: a long mail you want to read or answer fast.
Summarize the email below in 3 bullet points: what the sender wants, any deadline, and what I must do. Then suggest a reply of at most 4 sentences in a friendly, direct tone. If something is unclear, list it as a question instead of guessing.
Email:
"""
[PASTE EMAIL]
"""
Why it works: It names three items to extract, caps the reply length and bans guessing. Quote marks mark where the email starts and ends.
Simplify a text
Use it for: legal, technical or dense text you need in plain words.
Rewrite the text below for a reader with no background in [TOPIC]. Use short sentences and everyday words. Explain each technical term once in brackets. Keep all facts and numbers. Do not add anything new. Length: about half of the original.
Text:
"""
[PASTE TEXT]
"""
Why it works: It defines the reader, the style and the length, and it forbids new content, which cuts down on invented detail.
Minutes from notes
Use it for: turning rough meeting notes into a clean record.
Turn my notes into meeting minutes. Use these headings: Date and attendees, Decisions, Action items (task, owner, due date), Open questions. Use only what is in the notes. If an owner or date is missing, write "not set". Keep it under one page.
Notes:
"""
[PASTE NOTES]
"""
Why it works: Fixed headings give a stable format. "Use only what is in the notes" and "not set" stop the model from filling gaps.
Table from text
Use it for: pulling facts out of free text into rows and columns.
Read the text below and extract every [ITEM TYPE, e.g. product, person, offer] into a table with these columns: [COLUMN 1], [COLUMN 2], [COLUMN 3]. Use a Markdown table. Write "n/a" where the text gives no value. Do not guess. Add a final line with the number of rows.
Text:
"""
[PASTE TEXT]
"""
Why it works: You give the exact columns and a rule for missing values. The row count lets you spot a missed item.
Explain code
Use it for: understanding code you did not write.
Explain the code below to a developer who knows [LANGUAGE] but not this project. First give a 2-sentence summary of what it does. Then walk through it step by step in plain language. End with 3 risks or things that could break. Do not change the code.
Code:
"""
[PASTE CODE]
"""
Why it works: The reader level, the order (summary, steps, risks) and the "do not change" rule keep the answer focused.
Find a bug
Use it for: an error you cannot solve.
I get this error: [ERROR MESSAGE]. Expected behavior: [WHAT SHOULD HAPPEN]. Actual behavior: [WHAT HAPPENS]. Environment: [LANGUAGE, VERSION, OS]. Find the most likely cause. Give me the 3 best candidates ranked by likelihood, and for each a short test to confirm it. Then show the smallest fix for the top one.
Code:
"""
[PASTE THE RELEVANT CODE]
"""
Why it works: It gives the facts a debugger needs (error, expected, actual, environment) and asks for ranked causes with tests, not one confident guess.
Weigh a decision
Use it for: a choice with trade-offs.
Help me decide: [DECISION, e.g. rent or buy a laptop]. Options: [OPTION A], [OPTION B]. My priorities, most important first: [PRIORITY 1], [PRIORITY 2], [PRIORITY 3]. Facts I know: [FACTS]. List pros and cons for each option, point out what I may be missing, and ask me up to 3 questions that would change the answer. Then give a recommendation and say how sure you are.
Why it works: Ranked priorities let the model weigh trade-offs. The questions and the confidence statement show where the answer is shaky.
Plan a trip
Use it for: a first draft of an itinerary.
Plan a [NUMBER]-day trip to [PLACE] for [WHO, e.g. two adults]. Budget: about [AMOUNT] per day. Interests: [INTERESTS]. Pace: [RELAXED / BUSY]. Give a day-by-day plan with morning, afternoon and evening, travel time between stops, and one rainy-day backup. Mark anything I must book ahead. Say which details I should verify (opening hours, prices, transport) because you may be out of date.
Why it works: It sets budget, pace and format. The final line reminds you that opening hours and prices must be checked (knowledge cutoff).
Flashcards
Use it for: learning from a text or a topic.
Create 15 flashcards from the material below. Format each as "Q:" and "A:" on separate lines. Mix definitions, causes and examples. Keep each answer to 1 or 2 sentences. Start easy and end with harder cards. Use only the material.
Material:
"""
[PASTE NOTES OR TEXT]
"""
Why it works: A fixed Q/A format is easy to copy into a flashcard app. Mixed question types and rising difficulty help recall.
Translate with tone
Use it for: a translation that sounds right, not just correct.
Translate the text below from [SOURCE LANGUAGE] into [TARGET LANGUAGE]. Tone: [FORMAL / FRIENDLY / NEUTRAL]. Audience: [WHO]. Keep names, numbers and line breaks. Do not translate word for word; use natural phrasing. After the translation, list any phrase where you had to choose between two meanings.
Text:
"""
[PASTE TEXT]
"""
Why it works: Tone and audience change word choice. The list of doubtful phrases tells you where to check.
Feedback on a text
Use it for: an honest edit of your own writing.
Act as a careful editor. Review my text below, written for [AUDIENCE] with the goal of [GOAL]. Give feedback in three parts: 1) what works, 2) the 3 biggest problems, with a quote from my text for each, 3) a rewrite of the weakest paragraph. Be direct and specific. Do not rewrite the whole text.
Text:
"""
[PASTE TEXT]
"""
Why it works: A role, a goal and an audience give the model a standard to judge by. Asking for quotes keeps the feedback tied to the text.
Brainstorm ideas
Use it for: getting past a blank page.
Give me 12 ideas for [GOAL, e.g. a team event, a blog topic, a product name]. Context: [CONTEXT]. Limits: [BUDGET, TIME, RULES]. Group them as 4 safe ideas, 4 bold ideas and 4 odd ideas. One line each. Then pick your top 3 and say why in one sentence each.
Why it works: Grouping by risk forces variety. Limits keep ideas usable, and the top-3 step gives you a start.
How to build good prompts
Use five parts. You do not need all five every time.
- Role. Who should the model act as? "Act as a careful editor."
- Context. What does it need to know? Audience, background, facts, limits.
- Goal. What should the result do? State one clear task.
- Format. What should it look like? Bullets, table, length, headings.
- Examples. Show one sample of the style you want. A short example beats a long description.
Then add three habits:
- Say what to avoid. "Do not guess" and "use only the text" lower the risk of made-up facts.
- Mark your input. Put pasted text between quotes or tags so the model knows where it starts and ends.
- Iterate. If the first answer misses, say what is wrong and ask for a change. Short follow-ups often beat a rewrite of the whole prompt.
For apps and APIs, a system prompt holds the standing rules (role, tone, format) so you do not repeat them in every message. A lower temperature makes answers steadier in tools that let you set it (some newer models ignore it). If a task is hard, a stronger model may do better than a longer prompt; see how to choose a model.
Tip: Save prompts that work in a note. Swap the brackets each time.
Questions and answers
Do these prompts work in ChatGPT, Claude and Gemini?
Yes. They use plain language and no vendor tricks. Results differ by model, so adjust the wording if an answer misses the mark.
Why do the prompts use square brackets?
Brackets mark the parts you replace with your own text or facts, such as [TEXT] or [AUDIENCE]. Remove the brackets when you fill them in.
Can I trust the answers?
Not blindly. Models can state wrong facts with confidence. Check names, numbers, dates and code before you use them.
Is it safe to paste my own emails or notes?
Only if they hold nothing private and your provider terms allow it. Read what not to paste first, and remove names and numbers.
Read next
- Prompt
A prompt is the input you give an AI model: your question or task, plus any instructions, examples and text it should use.
- System prompt
A system prompt is a block of standing instructions, set by an app or developer, that shapes how a model behaves across a whole conversation.
- Hallucination
A hallucination is a fluent, confident AI answer that is false, invented or not backed by the sources it was given.
- What not to paste into AI chatbots
Passwords, client data, IDs and secrets do not belong in a chatbot: a traffic-light checklist, provider data rules, Swiss and EU basics, and safer options.
- How to choose an AI model
Match the model to the job: small for volume, mid for daily work, top or reasoning for hard problems and local for private data, with a decision tree and a test.
- Context window
The context window is the maximum amount of text, counted in tokens, that a model can read and write in one request.
- Temperature
Temperature is a setting that controls how random a model's word choices are: low gives steady answers, high gives more varied ones.