10 Claude Prompts for Story Writing That Unlock Your Best Work

Every writer knows the feeling: the idea is there, vivid and urgent, but the blank page pushes back. The character feels thin. The scene lacks tension. The opening does not hook. The plot meanders without stakes. These are not signs of a bad writer — they are signs of a writer who has not yet found the right question to ask. The right Claude prompts for story writing act like that question — a thinking partner that helps you excavate the story already inside the idea, rather than generating one from nothing.

Below are 10 prompt patterns for every stage of the story writing process — from the first spark of an idea to the final polish of a scene. Each includes a ready-to-use example, an explanation of why it works, and a tip for getting even more from it. Whether you write literary fiction, genre novels, short stories, or screenplays, these prompts are built to serve your voice, not replace it.

Claude is not a ghostwriter. The best use of Claude in a creative writing process is not to produce finished prose on your behalf — it is to think alongside you. To generate options you can react to. To surface questions you have not asked yet. To hold the structure while you focus on the language. Used this way, Claude does not dilute your voice. It helps you find it faster.

The prompts below are designed around that philosophy. They are not instructions to write your story for you. They are tools for generating raw material you can shape, questions that expose what your story is really about, and exercises that build the elements — character, world, tension, voice — that strong fiction depends on.

A story idea is not a story. A premise is not a plot. This prompt takes the raw spark of an idea and develops it into a concept with the four things every compelling story needs: a character with a want, an obstacle, a stakes, and a question the story will answer.

"I have a story idea: [describe your idea in 1-3 sentences]. Develop this into a story concept by identifying: (1) the protagonist and their core desire, (2) the central conflict or obstacle standing in their way, (3) the stakes — what they lose if they fail, (4) the central question the story will answer, and (5) the genre and tone that best fits this premise. Then suggest two alternative directions this concept could take that I may not have considered."

Why it works: Most story ideas fail at the execution stage because the concept was never fully developed before writing began. Separating desire, obstacle, stakes, and central question forces the clarity that makes plotting possible. The two alternative directions are especially valuable — they often reveal a version of the idea that is more original than the first instinct.

Flat characters are the most common reason readers disengage. This prompt builds a character from the inside out — starting with wound and want rather than appearance and occupation, which is where most character work goes wrong.

"Help me build a character for my story. Here is what I know so far: [describe what you have]. Develop this character by exploring: (1) their core wound — the formative experience that shaped how they see the world, (2) their conscious want vs. their unconscious need — what they think they want vs. what they actually need to grow, (3) their dominant flaw and how it will create conflict, (4) their voice — how they speak and think differently from other characters, (5) what they would never do — and what the story might force them to do anyway."

Why it works: The want vs. need distinction is one of the most powerful tools in character writing — it creates the gap that drives character arc. The "what they would never do" question is particularly generative: whatever the answer is, that is often the thing your story needs to put them through. These prompts reveal internal structure, which is what separates memorable characters from functional ones.

World-building is not about creating encyclopedias. It is about building a world with enough internal logic and sensory texture that readers feel they have entered it. This prompt focuses world-building on the details that actually affect story and character — not the details that fill notebooks but never make it to the page.

"I am building the world for a [genre] story set in [brief description of the setting]. Help me develop: (1) the single most important rule or law of this world — the thing that shapes every character's life, (2) what ordinary people fear most in this world and why, (3) three sensory details that make this world feel distinct and lived-in, (4) one thing that seems normal here but would horrify an outsider, (5) how the protagonist's specific background shapes how they experience this world differently from others."

Why it works: Grounding world-building in what people fear and what they consider normal creates a world with emotional texture, not just geography. The final question — how the protagonist experiences this world differently — is the bridge between world-building and character work, which is where the best stories live.

The opening line of a scene sets the energy for everything that follows. When you are staring at a blank page knowing what needs to happen in a scene but not how to begin it, this prompt generates multiple entry points so you can choose the one that feels right.

"I need to write a scene where [describe what happens in the scene]. The emotional tone I want is [describe the feeling]. The character's internal state going into this scene is [describe]. Write 5 different opening lines for this scene, each using a different technique: (1) in medias res action, (2) a sensory detail that sets the mood, (3) a line of dialogue that creates immediate tension, (4) an internal thought that reveals character, (5) an unexpected observation that earns a second sentence."

Why it works: Different opening techniques create different emotional contracts with the reader. Seeing five options side by side makes it easy to identify which one matches your instinct for the scene — and sometimes the act of reading them triggers an entirely different opening you would not have reached without the prompt. Even one good opening line is often enough to unlock the rest of the scene.

Dialogue that sounds natural in your head often reads flat on the page. This prompt diagnoses what is not working in a dialogue exchange and rewrites it to carry subtext, reveal character, and move the scene forward simultaneously.

"Here is a dialogue exchange from my story: [paste the dialogue]. The two characters are [describe their relationship and current tension]. What each character wants from this conversation: Character A wants [X], Character B wants [Y]. Rewrite this dialogue so that: (1) neither character says exactly what they mean, (2) the subtext of each character's real agenda is clear to the reader but not to the other character, (3) the power dynamic shifts at least once, (4) each character's voice is distinct. Keep the same basic information being exchanged."

Why it works: The instruction that neither character says exactly what they mean is the single most important rule in dialogue writing. Real conversations — especially charged ones — are never about what they appear to be about. Asking Claude to preserve the information while adding subtext teaches the craft principle by showing the before and after side by side.

Every writer hits the point where the story stops working and they cannot see why. This prompt diagnoses structural problems and generates multiple solutions so you can choose the path that fits your vision rather than taking the first fix that appears.

"My story has a plot problem. Here is what I have so far: [describe your story up to the problem point]. The problem is: [describe exactly where and how the story is breaking down — the character's motivation feels unclear, the plot requires a coincidence, the pacing collapses in the middle, etc.]. Diagnose the root cause of this problem — is it a character issue, a structural issue, or a logic issue? Then give me three different solutions, each taking the story in a meaningfully different direction, and explain the trade-offs of each."

Why it works: Asking Claude to diagnose the root cause before offering solutions prevents surface-level fixes that do not address the real problem. Most plot problems are character problems in disguise: if a character would not believably make a required decision, the fix is not a new plot event — it is building the character motivation earlier. The three-solution format with trade-offs lets you choose with full information rather than taking the first plausible answer.

Tension is not conflict. Conflict is what happens. Tension is the feeling that something could happen — and the reader's dread of or desire for it. This prompt is designed for scenes that have the right events but feel flat because the tension has not been layered in.

"Here is a scene from my story that feels flat or low-stakes: [paste the scene or describe it in detail]. The emotional tension I want the reader to feel is [describe: dread, anticipation, heartbreak, unease, excitement]. Identify three specific moments in this scene where tension could be heightened without changing any plot events. For each moment, explain the technique — withholding information, dramatic irony, physical sensation, the character noticing the wrong thing — and rewrite that moment using it."

Why it works: The constraint of not changing any plot events is what makes this prompt useful rather than just generating a rewrite. It teaches you to find the tension already latent in your scene rather than adding new events. Naming the technique alongside the rewrite makes the learning explicit, so you can apply it to the next scene without the prompt.

Voice is the hardest element of fiction to develop and the easiest to lose when you are writing fast or uncertain. This prompt helps you identify what your prose voice is doing well, what is inconsistent, and how to deepen it across a piece.

"Here is a passage from my story: [paste 200-400 words of your writing]. Analyse the voice by identifying: (1) three things this prose is doing distinctively well, (2) one recurring habit that is weakening the voice — filter words, passive constructions, over-explanation, hedging, over-punctuation, (3) the emotional register this voice is working in, and (4) one writer whose style has some kinship with this passage and why. Then rewrite one paragraph to intensify what the voice is already doing well, without changing the content."

Why it works: Starting with what the voice is doing well before identifying weaknesses produces more useful feedback than pure critique — and it often reveals stylistic tendencies you were not consciously aware of. The comparative author suggestion is not to imitate but to orient — knowing that your prose has kinship with a particular writer gives you a library of technique to study deliberately.

Writer's block is rarely a lack of ideas. It is usually fear — of writing badly, of making the wrong choice, of the story not living up to the idea. This prompt bypasses the paralysis by generating low-stakes raw material you are not attached to.

"I am stuck on my story and cannot move forward. Here is where I am: [describe the story and where you have stopped]. I think I am stuck because [honest description: I do not know what happens next / the scene feels wrong / I have lost confidence in the story / the character's motivation is unclear]. Give me five wildly different ways the story could move forward from this point — ranging from the most expected to the most unexpected. Do not worry about whether they fit my current plan. I just need options to react to."

Why it works: The instruction to range from expected to unexpected is key — the unexpected options are often wrong for your story but right for unsticking you. Reacting to options you do not like is often what reveals what you actually want. The invitation to generate options you are not attached to removes the pressure of commitment, which is usually what is causing the block in the first place.

Before you show your work to a beta reader or editor, it is worth running it through a rigorous structural critique. This prompt turns Claude into a tough but specific reader who evaluates your story at the macro level — structure, character arc, pacing, theme — before you invest time in line-level revision.

"Act as an experienced fiction editor. Review the following story summary or outline: [paste your story summary, outline, or synopsis]. Score it out of 10 on: premise originality, protagonist depth, central conflict clarity, pacing and structure, and thematic coherence. For each category, identify the single biggest weakness. Then identify the two structural problems most likely to cause a reader to put the story down, and suggest one specific fix for each. Be direct — do not soften the feedback."

Why it works: Structural problems found at the outline stage cost an hour to fix. The same problems found after the first draft is complete can cost weeks. The five-category scoring framework forces specificity rather than general impressions. Asking Claude to identify the two problems most likely to cause a reader to disengage is more useful than a ranked list of issues — it prioritises ruthlessly, which is what editing requires.

The most important thing to remember when using Claude for creative writing is that the output is raw material, not finished work. Use Claude's responses as a springboard — take what resonates, discard what does not, and always rewrite in your own voice. The goal is not to speed up the writing. It is to deepen the thinking so the writing can be more fully yours.

Save the prompts that work best for your process as reusable templates in Chat Smith so you can deploy them at the right moment in your writing workflow without rebuilding them from scratch. The Character Builder prompt at the start of a new project. The Scene Starter when a chapter will not open. The Writer's Block Breaker when you have been staring at the same paragraph for an hour. Each prompt becomes more valuable the more precisely you know when to use it.

Using these prompts steers you away from the structural failures that derail most first drafts. Stories that begin without a fully developed premise collapse in the middle because the concept was never stress-tested. Characters built from the outside in — appearance and job first — feel like vessels for plot rather than people driving it. Scenes that have the right events but no layered tension lose readers even when nothing is technically wrong. Dialogue that says what it means reads like exposition in quotation marks.

Each prompt in this guide targets one of these failure modes directly. The Concept Developer addresses underdeveloped premises. The Character Builder addresses characters built from the outside in. The Tension Escalator addresses scenes that have events but not feeling. The Dialogue Sharpener addresses characters who say what they mean. The pattern is always the same: build the internal structure first, and the surface will follow.

Great stories are not written — they are rewritten, rethought, and rebuilt from stronger foundations. These 10 Claude prompts for story writing give you the tools to build those foundations faster and more deliberately than working in isolation. Use them to develop the ideas that matter to you, solve the problems that are blocking you, and sharpen the work you are proudest of. The story you are trying to write is already there. These prompts help you find it.

A story writing workflow involves many different kinds of thinking — concept development, character work, scene-level craft, structural review — and each requires a different kind of prompt. Keeping all of those prompts organised and immediately deployable is exactly where Chat Smith comes in. Chat Smith is an all-in-one AI platform that lets you save every story writing prompt as a reusable template, organise them by writing stage or project, and launch any prompt in one click across Claude, GPT, Gemini, and other leading models.

Instead of rebuilding your character prompt from scratch at the start of every new project, or hunting for your tension escalation template when a scene goes flat, Chat Smith gives you a clean, searchable library of your best-performing prompts. You can run the same prompt across multiple models to compare the responses, build a consistent creative workflow across every story you write, and spend less time setting up the prompts and more time doing the actual writing.

1. Will using Claude make my writing sound like AI?

Not if you use it as a thinking tool rather than a writing tool. The prompts in this guide are designed to generate raw material, options, and structural thinking — not finished prose to copy. When you take Claude's output, react to it, discard what does not fit, and rewrite what does in your own language, the result is entirely your voice. The risk of AI-sounding prose only arises when output is used without transformation.

2. Can these prompts work for short stories as well as novels?

Yes. The prompts are built around story fundamentals — character, conflict, tension, voice — that apply regardless of length. For short stories, the Concept Developer and Scene Starter prompts are especially useful because the margin for structural error is smaller. For novels, the Plot Problem Solver and Structural Critiquer become more essential as the architecture gets more complex.

3. Can I use these prompts for genre fiction — fantasy, thriller, romance?

Absolutely. The prompts are structured around universal story principles that apply across all genres. The World-Building prompt is particularly useful for fantasy and science fiction. The Tension Escalator is essential for thriller and horror. The Dialogue Sharpener works especially well for romance, where subtext carries enormous weight. Specify your genre in the prompt and Claude will calibrate accordingly.

4. What should I do if Claude's output does not match my vision for the story?

That is useful information. When Claude's output does not match your vision, it often means your vision is clearer than you realised — you just needed something wrong to react against. Note what specifically did not fit and why, then refine your prompt with that information. The mismatch is part of the process, not a failure of it. The best prompts are usually the third or fourth iteration, not the first.