Claude Prompts Library
A curated library of 15 high-value Claude prompts, organized by category, writing, coding, analysis, research, and business. Every one is a real copy-paste prompt with a short usage note.
Last updated June 21, 2026
How to use this library
Each prompt below is ready to paste into Claude. Replace anything in [brackets] with your details, and where the prompt expects source material, paste your document, data, or code directly so Claude works from real context. The usage note under each prompt tells you when it shines and how to adapt it. These prompts lean on what Claude does well, following structure, holding long context, and respecting constraints, so the more specific you are about audience, tone, length, and what to avoid, the better the result.
Writing & communication
1. Sharpen any draft. "Improve this [email/post/paragraph] for clarity and impact. Keep my meaning, cut filler, and make it [tone]. Show the revised version, then a one-line note on what you changed: [paste text]"
Use when: you have a rough draft and want a tighter version plus a quick rationale.
2. Adapt for a specific audience. "Rewrite this for [audience, e.g. busy executives / non-technical customers]. Adjust vocabulary and length, lead with what they care about, and keep it under [N] words: [paste text]"
Use when: the same message needs different framing for different readers.
3. Outline before you write. "Act as an editor. Give me a structured outline for a [article/report] on [topic] for [audience], with section headers and one sentence on what each section should cover."
Use when: you want a plan you can fill in rather than a finished draft.
Coding & technical
4. Explain code in plain English. "Explain what this code does step by step, call out anything risky or non-obvious, and note one thing you would improve: [paste code]"
Use when: you inherited code or want a second read before changing it.
5. Debug with reasoning. "Here is my code and the error / unexpected behavior. Think through the likely causes, identify the most probable one, and give me a corrected version with a short explanation: [paste code + error]"
Use when: something is broken and you want a reasoned fix, not a guess.
6. Write tests. "Write [framework] unit tests for this function. Cover normal cases, edge cases, and failure cases, and briefly say what each test verifies: [paste function]"
Use when: you need coverage fast and want the reasoning behind each case.
Analysis & reasoning
7. Summarize a long document. "Summarize this into [N] bullet points for [audience], focusing on [decisions / risks / action items]. Then add a one-line bottom line: [paste document]"
Use when: you need the signal from a long report quickly.
8. Pressure-test an idea. "Act as a thoughtful skeptic. Here is my plan: [describe]. Give me the three strongest objections, the assumptions I am making, and what would have to be true for this to work."
Use when: you want honest pushback before committing.
9. Compare options in a table. "Compare [option A], [option B], and [option C] for [my goal]. Output a table of pros, cons, cost, and best-fit use case, then a one-line recommendation with the trade-off."
Use when: you are choosing between tools, vendors, or approaches.
10. Extract structure from messy text. "Pull the key data from this text into a clean table with columns [list columns]. If a value is missing, mark it 'N/A' rather than guessing: [paste text]"
Use when: you have unstructured notes and need organized data.
Research & learning
11. Explain a hard concept. "Explain [concept] to me as if I know [my level]. Use one concrete analogy and one real example, then list two common misconceptions."
Use when: you are learning something new and want it to actually stick.
12. Build a study plan. "I want to learn [topic] in [timeframe], starting from [current level]. Create a week-by-week plan with goals, what to practice, and a simple way to check my progress each week."
Use when: you want structure for self-directed learning.
13. Interrogate a topic. "I am researching [topic]. Ask me five sharp questions that would expose gaps in my understanding, then, based on my answers, tell me what to dig into next."
Use when: you want Claude to guide your research rather than just answer.
Business & productivity
14. Turn notes into a clear plan. "Here are my messy notes from [meeting/idea]. Organize them into a short plan: objective, key decisions, action items with owners, and open questions: [paste notes]"
Use when: you need to convert scattered thoughts into next steps.
15. Draft a first version of anything. "Draft a [proposal / job description / FAQ / policy] for [context]. Keep it [tone], under [N] words, and leave [bracketed placeholders] where I need to add specifics."
Use when: a blank page is the bottleneck and you just need a strong starting draft.
FAQ
What makes a prompt work well with Claude specifically?
Claude responds well to clear structure, explicit roles, and context up front. It handles long inputs and multi-step instructions reliably, so you can give it a role, the material to work from, the task, and the output format in one prompt. It is also strong at following constraints, word counts, tone, what to avoid, and at saying when it is unsure rather than guessing. The prompts in this library are written to play to those strengths: each sets context, names the task, and specifies the format you want back.
How is this library different from a generic list of prompts?
A generic list dumps dozens of one-line ideas. This library is organized by category, writing, coding, analysis, research, and business, and every entry is a complete, copy-paste prompt with bracketed variables to fill in plus a short usage note explaining when and how to use it. The goal is fewer, higher-value prompts you can actually run and adapt, rather than a long undifferentiated feed. Think of it as a starting toolkit you customize, not a catalog to scroll.
How do I customize these prompts?
Replace anything in [brackets] with your specifics, then add one or two lines of constraints, audience, tone, length, and what to avoid. If the first answer is close but not right, refine with a follow-up like 'make it shorter,' 'use a more formal tone,' or 'add a counterargument.' For complex work, paste your source material (a document, data, or code) into the prompt so Claude works from real context instead of assumptions. Iterating with follow-ups is usually faster than writing one perfect mega-prompt.
Do these prompts work on the free version of Claude?
Yes. These are plain-text prompts, so they work in the Claude app and on the web regardless of plan; the difference between free and paid tiers is mainly usage limits and access to the most capable models, not whether a given prompt format works. For long or complex tasks you may hit limits faster on a free tier, and the most advanced model may produce stronger results, but the prompt structures here are model-agnostic.
Can I use these prompts for work and commercial projects?
Generally yes, prompts are instructions, and the output is yours to use subject to Anthropic's usage policies and your own organization's rules. For commercial or sensitive work, review your company's AI policy, avoid pasting confidential data into consumer tools, and always fact-check and edit the output before it ships. Treat Claude as a fast, capable drafting partner whose work you review, not as a final authority.
How do I get better, more reliable answers from Claude?
Give context before the task, be explicit about the output format, and set constraints. Ask Claude to think through a problem step by step for analysis and reasoning tasks, and to show its assumptions. Provide an example of the style or structure you want when it matters. If accuracy is critical, ask it to flag anything it is unsure about and to cite where a claim would normally come from. Finally, iterate, the second or third refined prompt almost always beats the first.
Related: Claude prompts.