ChatGPT Prompts for Writing
AI writing assistants are only as good as the prompts that guide them. Whether you are drafting blog posts, editing essays, brainstorming creative fiction, or repurposing existing content into new formats, the structure and specificity of your prompt determines whether you get something publishable or something that reads like a template. The prompts below cover the full spectrum of writing tasks and are designed to produce output that needs minimal editing.
For blog posts, specify the target audience, expertise level, desired length, and whether you want a conversational or authoritative tone. Include the key points you want covered rather than letting the AI decide the structure. For essays and long-form pieces, provide a thesis or argument direction and indicate whether you want the AI to present multiple perspectives or advocate a position. Creative writing prompts benefit from setting the scene, defining character constraints, and specifying the emotional arc you want — the more specific your creative direction, the less generic the output.
Editing prompts are often more valuable than generation prompts. Ask the AI to identify weak transitions, flag passive voice, tighten wordy sentences, or adjust reading level — but do one task at a time rather than asking for a general "improve this." Tone adjustment prompts should include a sample of the desired voice or reference a specific author or publication style. The writers getting the most from AI treat it as a collaborative tool with a defined role in their workflow, not a replacement for the creative process.
8 Ready-to-Use Writing Prompts
Copy any prompt, replace the {{variables}} with your details, and paste into ChatGPT, Claude, or any AI tool.
Blog Post Draft
Write a blog post about "{{topic}}" for {{target_audience}}. Length: {{word_count}} words Tone: {{tone}} (e.g., conversational and practical, authoritative and research-backed, casual and opinionated) Expertise level of reader: {{expertise_level}} (beginner / intermediate / expert) Structure: - **Title**: Specific and benefit-driven. No clickbait. - **Opening paragraph**: Start with a concrete scenario, surprising stat, or direct statement the reader will nod along to. Do NOT start with "In today's world..." or any throat-clearing. - **Body**: {{section_count}} main sections with H2 headings. Each section should make one clear point supported by an example, analogy, or data point. - **Conclusion**: Summarize the key takeaway in 2 sentences and end with a specific next step the reader can take today. Key points to cover: {{key_points}} Writing rules: - Use short paragraphs (3-4 sentences max) - Include at least one concrete example per section - Avoid filler phrases: "it's important to note", "it goes without saying", "at the end of the day" - Write in active voice. Limit sentences to 20 words where possible.
Why it works: Bans common AI filler phrases, enforces paragraph length and sentence structure, and requires concrete examples -- producing drafts that read like a human wrote them rather than obvious AI output.
Essay Editor
Edit the following essay for clarity, structure, and persuasiveness. Do not rewrite it from scratch -- preserve my voice and arguments while improving the execution. My essay: """ {{essay_text}} """ Editing pass instructions (perform each in order): 1. **Structure check**: Is the thesis clear within the first two paragraphs? Does each paragraph advance the argument? Flag any paragraph that could be cut or merged. 2. **Transition audit**: Identify the 3 weakest transitions between paragraphs. Suggest specific replacement transitions. 3. **Sentence-level editing**: - Flag and fix passive voice (unless it is genuinely better in context) - Cut filler words and redundant phrases - Break any sentence over 30 words into two 4. **Argument strength**: Identify the single weakest claim and suggest how to strengthen it with evidence, a counterargument, or sharper reasoning. 5. **Reading level**: Current essay targets {{reading_level}} audience. Adjust vocabulary and sentence complexity if any sections drift from that target. Output format: Return the edited essay with your changes tracked using **[CHANGED: reason]** annotations inline so I can review each edit.
Why it works: Gives the AI a specific editing methodology to follow in sequence rather than asking for vague "improvement." The inline annotation format lets you accept or reject each change individually.
Creative Fiction Scene
Write a scene for a {{genre}} story. Setting: {{setting}} Time period: {{time_period}} Point of view: {{pov}} (first person / third person limited / third person omniscient) Tone: {{tone}} (e.g., tense and atmospheric, darkly humorous, lyrical, spare and Hemingway-esque) Characters in this scene: - {{character_1}}: {{character_1_description}} - {{character_2}}: {{character_2_description}} Scene purpose: {{scene_purpose}} (e.g., reveal the betrayal, build romantic tension, establish the world, show the protagonist's flaw) Constraints: - Open with action or dialogue, not description - Show character emotions through physical behavior and dialogue subtext, not internal narration stating feelings directly - Include at least one specific sensory detail (smell, texture, sound) per paragraph - The scene should end on a beat of tension, unresolved question, or emotional shift -- not a neat resolution - Length: {{scene_length}} words Do not: use purple prose, adverb-heavy dialogue tags (she said angrily), or cliched metaphors (heart pounding, blood running cold). Find fresh language.
Why it works: Enforces show-don't-tell principles, bans common creative writing cliches, and requires specific sensory details -- the exact notes a writing workshop would give.
Email Rewrite (Tone Adjustment)
Rewrite the following email to change its tone from {{current_tone}} to {{target_tone}}. Original email: """ {{email_text}} """ Requirements: - Preserve the core message, all factual content, and the call-to-action - Adjust vocabulary, sentence length, and formality level to match the target tone - If the target tone is more professional: remove contractions, hedging language, and casual asides - If the target tone is more casual: add contractions, shorter sentences, and a warmer sign-off - If the target tone is more assertive: lead with the ask, cut qualifiers ("just", "maybe", "I think"), and use direct statements - If the target tone is more diplomatic: add acknowledgments, soften directives into questions, and provide reasoning before requests Also provide: 1. A 1-sentence summary of the key changes you made 2. A "before/after" comparison of the single most impactful sentence change Keep the same approximate length as the original. Do not add information that was not in the original email.
Why it works: Provides tone-specific transformation rules rather than leaving "tone" as a vague concept. The before/after comparison teaches you to internalize the adjustments for future emails.
Newsletter Intro
Write the opening section for a {{frequency}} newsletter called "{{newsletter_name}}" about {{topic}}. Audience: {{audience}} This issue's theme: {{theme}} Key story or insight to highlight: {{key_insight}} Tone: {{tone}} (e.g., witty and informed, warm and conversational, sharp and opinionated) Structure: - **Subject line**: Under 50 characters. Curiosity or specificity -- never both at once. - **Preview text**: Under 90 characters. Complements the subject line (does not repeat it). - **Opening hook** (2-3 sentences): Pull the reader in with a surprising fact, a bold claim, a relatable moment, or a quick story. Do NOT start with "Welcome to this week's edition of..." - **Transition to theme** (1-2 sentences): Connect the hook to this issue's main theme - **Table of contents teaser**: 3-4 bullet points previewing what is in the issue, written as intriguing one-liners (not boring summaries) Newsletter sections to tease: {{sections}} The intro should make the reader want to keep scrolling. Write it in under 150 words.
Why it works: Separates subject line, preview text, and body as distinct writing tasks with their own constraints. The "no welcome to this edition" ban forces a hook that earns attention instead of assuming it.
Technical Explainer
Explain {{technical_concept}} for an audience of {{audience}}. Their existing knowledge: {{existing_knowledge}} Goal of this explanation: {{goal}} (e.g., help them evaluate whether to adopt it, teach them to implement it, give a conceptual overview for a presentation) Format: {{format}} (e.g., blog post section, documentation page, presentation script, internal memo) Length: {{word_count}} words Writing instructions: 1. Start with a one-sentence plain-English definition. No jargon in the first sentence. 2. Explain WHY this concept matters before explaining HOW it works. Motivation before mechanism. 3. Use a concrete analogy from everyday life to bridge the gap from unfamiliar to familiar. 4. Walk through a real example: show a simplified {{example_type}} (e.g., code snippet, workflow, before/after comparison) and annotate each step. 5. Address the #1 misconception about {{technical_concept}} and correct it. 6. End with: "In practice, this means..." followed by a concrete implication for the reader's work. Do not: assume knowledge the audience does not have, use acronyms without defining them, or sacrifice accuracy for simplicity -- find explanations that are both correct and accessible.
Why it works: Follows the pedagogy pattern of definition, motivation, analogy, example, misconception, and implication -- which mirrors how effective technical educators actually structure explanations.
Content Repurposing: Blog to Tweet Thread
Convert the following blog post into a Twitter/X thread of {{thread_length}} tweets. Blog post: """ {{blog_content}} """ Thread rules: - **Tweet 1 (hook)**: A bold, standalone statement or question that makes someone stop scrolling. Must work without any context. End with "A thread." or similar. - **Tweets 2 through {{thread_length_minus_1}}**: One idea per tweet. Each tweet should be self-contained enough to make sense if someone screenshots just that one. - Use specific numbers, examples, or frameworks from the blog post - No tweet should start with "And" or "Also" or "Additionally" - Alternate between: insight tweets, example tweets, and contrarian/surprising takes - **Final tweet**: Summarize the core lesson in one sentence + a CTA (follow, bookmark, or link to the full post) Formatting: - Max 280 characters per tweet - Use line breaks within tweets for readability - No hashtags (they reduce engagement on threads) - No emojis unless they replace a word and add clarity - Number each tweet: 1/, 2/, etc. Also provide: 3 alternative hook tweets I can A/B test for Tweet 1.
Why it works: Treats each tweet as an independent unit (critical for engagement), bans weak transitions, and requests alternative hooks -- the same workflow professional content creators use when repurposing.
Writing Style Analysis
Analyze the writing style of the following text and produce a reusable style guide I can use to prompt AI tools to write in this voice.
Text sample:
"""
{{text_sample}}
"""
Analyze and document:
1. **Voice & Tone**: Is it formal or casual? Warm or detached? Authoritative or exploratory? Use specific adjectives and cite example sentences.
2. **Sentence structure**: Average sentence length (short/medium/long). Does the writer vary length for rhythm? Use fragments? Favor simple or compound-complex sentences?
3. **Vocabulary level**: Is the language plain, elevated, technical, or colloquial? Any distinctive word choices or pet phrases?
4. **Paragraph structure**: Short punchy paragraphs or longer developed ones? How does the writer transition between ideas?
5. **Rhetorical devices**: Does the writer use metaphors, analogies, rhetorical questions, lists, repetition, humor, or direct address? Give examples.
6. **Point of view & perspective**: First person, second person, third person? Does the writer include personal anecdotes or stay abstract?
7. **What the writer avoids**: Note anything conspicuously absent (e.g., no jargon, no passive voice, no hedging language, no exclamation marks).
Finally, produce a **"Write Like This" prompt block** -- a 5-8 sentence instruction paragraph I can paste into any AI tool to replicate this style. It should be specific enough that the AI output would be difficult to distinguish from the original author.Why it works: Goes beyond vague "analyze the tone" requests by examining seven specific dimensions of style and producing a reusable prompt block -- turning a one-time analysis into a permanent tool.
Recommended tools & resources
Browse writing prompt templates from the community library.
Prompt BuilderBuild custom writing prompts with guided structure.
Prompt TipsTechniques for getting better writing outputs from AI.
Prompt PatternsReusable prompt structures for different writing formats.
PersonasDefine writer personas for consistent voice and style.
Prompt ScoreEvaluate and refine your writing prompts before using them.