0Pricing
AI Prompt Engineering · Lesson

Creative and Storytelling Prompts

Fiction, scripts, poems — character, conflict, and narrative arc prompts.

Creative and Storytelling Prompts is a free AI Prompt Engineering lesson on CoddyKit — lesson 4 of 4. You can read the complete lesson below for free — then practise it hands-on in the browser with a built-in code editor and a 24/7 AI tutor. It is part of the AI Prompt Engineering learning path, one of 4 lessons in the course, and your progress syncs across the web and the CoddyKit app.

Creative Prompting: More Than 'Be Creative'

Creative writing prompts require a different approach than informational or professional writing prompts. Vague creative prompts — "write a story about friendship" — produce generic, forgettable output because the model defaults to the most common patterns in its training data.

Counterintuitively, more specific creative prompts produce more original output. Constraints force the model away from defaults and toward combinations that are distinctive. The creative elements you specify — character, setting, conflict, POV, genre, tone — each narrow the solution space toward something specific and interesting.

Fiction Prompts: The Character Specification

Character is the foundation of compelling fiction. A well-specified character in a prompt produces more dimensional writing than a generic character placeholder. Specify:

  • Name: establishes nationality, era, feel
  • Motivation: what does this character want above all else?
  • Flaw: what internal weakness prevents them from getting it easily?
  • Voice trait: how do they speak? (clipped, verbose, formal, evasive)

The flaw is the most important element — it creates internal conflict and drives the story. A character without a flaw has nothing to overcome.

Fiction Prompts: Setting and Conflict

Setting and conflict are two more essential fiction prompt elements:

Setting is not just location — it is location + time + atmosphere. "A detective's office" is location. "A detective's office on the last night of Prohibition, rain hammering the window, one lamp working" is setting.

Conflict is the engine of narrative. Specify the conflict type explicitly:

  • Person vs person
  • Person vs themselves (internal)
  • Person vs nature or environment
  • Person vs society or system

Without a specified conflict, the model produces slice-of-life scenes, not stories.

Fiction Prompt: POV and Genre

Two more fiction elements that dramatically affect output:

Point of view (POV):

  • First person: intimate, limited to one consciousness, creates identification
  • Third person limited: follows one character but with narrative distance
  • Third person omniscient: access to all characters' thoughts, epic scope
  • Second person: unusual, creates direct reader participation

Genre: each genre has conventions the model applies when named — noir, literary fiction, horror, magical realism, sci-fi, thriller. Naming the genre is a shorthand for dozens of conventions the model already knows.

A Complete Fiction Prompt

Combining all fiction elements produces a highly specified prompt that generates distinctive output:

import openai

client = openai.OpenAI(api_key='sk-...')

def generate_fiction(character_name, motivation, flaw, setting, conflict,
                     pov, genre, length_words, tone):

    prompt = (
        f'Write a short story in the {genre} genre.\n'
        f'Character: {character_name}. Motivation: {motivation}. Flaw: {flaw}.\n'
        f'Setting: {setting}\n'
        f'Conflict: {conflict}\n'
        f'Point of view: {pov}.\n'
        f'Tone: {tone}.\n'
        f'Length: approximately {length_words} words.\n'
        'Show, do not tell. No dialogue tags except "said". '
        'Begin in medias res — in the middle of action, not backstory.'
    )

    response = client.chat.completions.create(
        model='gpt-4o-mini',
        messages=[{'role': 'user', 'content': prompt}]
    )
    return response.choices[0].message.content

print(generate_fiction(
    character_name='Nadia Solis',
    motivation='to prove she deserves her late father\'s restaurant',
    flaw='she sabotages close relationships whenever she feels vulnerable',
    setting='a failing Havana-style restaurant in Miami, midnight before a health inspection',
    conflict='person vs themselves — the inspector is her estranged sister',
    pov='third person limited, close to Nadia',
    genre='literary fiction',
    length_words=600,
    tone='tense, bittersweet'
))

Poetry Prompts: Form Specification

Poetry prompts need form specification. Without it, the model defaults to free verse — which is valid but may not be what you want. Common poetry forms to specify:

  • Haiku: 5-7-5 syllable structure, three lines, nature imagery or juxtaposition
  • Sonnet: 14 lines, iambic pentameter, Shakespearean (ABAB CDCD EFEF GG) or Petrarchan
  • Villanelle: 19 lines, two repeating rhymes, two refrains (Dylan Thomas's 'Do Not Go Gentle' is the famous example)
  • Free verse: no formal constraints — specify line length and lineation instead
  • Prose poem: paragraph form but with lyrical density and compression

Poetry Prompts: Theme and Mood

Beyond form, poetry prompts benefit from specifying theme, central image, and mood:

"Write a Petrarchan sonnet. Theme: the way familiar places feel different after grief. Central image: the light in a kitchen window at dusk. Mood: elegy without sentimentality — mourning that has not tipped into self-pity. Do not mention death directly. Use concrete sensory images rather than abstractions."

The "concrete sensory images rather than abstractions" instruction is particularly powerful — it prevents the model from reaching for vague emotional language ("deep sorrow") instead of specific images that create emotional resonance.

Iterative Creative Generation

Creative output benefits from an iterative approach. Instead of one definitive generation, generate multiple variations and select the most promising elements:

import openai

client = openai.OpenAI(api_key='sk-...')

def generate_story_openings(premise, genre, num_variations=3):
    prompt = (
        f'Write {num_variations} different opening paragraphs (100 words each) '
        f'for a {genre} story about: {premise}\n'
        'Each opening should:\n'
        '  - Begin differently (different moment, angle, or technique)\n'
        '  - Establish a distinct tone\n'
        '  - Create immediate tension or intrigue\n'
        'Number them: Opening 1:, Opening 2:, Opening 3:\n'
        'Do not explain the openings — just write them. '
        'After all three, add one sentence noting the key difference in approach for each.'
    )

    response = client.chat.completions.create(
        model='gpt-4o-mini',
        messages=[{'role': 'user', 'content': prompt}]
    )
    return response.choices[0].message.content

print(generate_story_openings(
    premise='a forensic accountant discovers her firm has been laundering money for her own family',
    genre='thriller'
))

Creative Constraints as Creativity Drivers

Constraints in creative writing are not limitations — they are creativity drivers. The most interesting creative prompts often include unexpected constraints that force original solutions:

  • "Write this scene using only dialogue. No narration, no action beats."
  • "Tell the story from the perspective of an inanimate object in the room."
  • "Write the same moment from three different characters' perspectives, each in 100 words."
  • "Write a love story in which neither character ever speaks — only actions and observations."

Unusual constraints are one of the most effective techniques for generating creative output that differs from the model's default patterns.

World-Building Prompts

For speculative fiction (sci-fi, fantasy), world-building prompts need specific parameters to avoid the model producing generic fantasy or sci-fi tropes:

  • The central rule: what is the one thing about this world that is different from ours?
  • The cost of the rule: every magical or sci-fi element should have a cost or complication
  • What is preserved: what about the world is deliberately familiar, to anchor readers?
  • Tone of the world: gritty and realistic? Hopeful? Satirical?

"In this world, memories can be transferred between people — but each transfer erases part of the donor's own memory. Show a memory-transfer clinic in a city where this is routine."

Avoiding Generic Creative Output

AI models default to common creative patterns. These are the signals of generic creative output to watch for and correct:

  • Protagonist who has lost someone close and is now on a journey of healing
  • Descriptions of sunsets, rain, or coffee as emotional metaphors
  • Stories that end with the protagonist learning a lesson and growing
  • Poetry that rhymes in an obvious sing-song pattern without depth
  • Dialogue that exists only to explain backstory

Counteraction: specify the unexpected. "The protagonist does not grow or learn — they make the same mistake again at the end, but we understand it differently now."

Knowledge Check: Creative Prompts

You want to write a short horror story (500 words) about a character named Dr. Vance who is a marine biologist. You want it to be genuinely unsettling, not generic horror. Which prompt produces the most distinctive output?

Recap: Creative and Storytelling Prompts

Creative prompts produce more original output when they are more specific, not less. The key fiction elements to specify are: character (name, motivation, flaw), setting (location + time + atmosphere), conflict type, POV, genre, length, and tone. Poetry prompts need form specification (haiku, sonnet, villanelle, free verse) plus theme, central image, and mood.

Unusual constraints — unexpected POVs, show-don't-tell rules, restrictions on common tropes — are the most effective technique for generating creative output that differs from the model's defaults. Iterative generation (multiple variations, then selection) produces better creative output than one-shot generation.

Frequently asked questions

Is the “Creative and Storytelling Prompts” lesson free?

Yes — the full text of “Creative and Storytelling Prompts” is free to read here on the web, and the AI Prompt Engineering course includes 4 lessons in total. To practise it interactively (a built-in code editor and a 24/7 AI tutor) and unlock the rest of the AI Prompt Engineering course, upgrade to CoddyKit PRO.

What will I learn in “Creative and Storytelling Prompts”?

Fiction, scripts, poems — character, conflict, and narrative arc prompts. You practise AI Prompt Engineering with hands-on code you run directly in the browser, and a 24/7 AI tutor answers your questions as you work through the lesson.

Do I need any experience to start AI Prompt Engineering?

No prior experience is required. AI Prompt Engineering on CoddyKit is structured for beginners through advanced learners; this is — lesson 4 of 4, so you can start here or from the beginning and move at your own pace.

How long does the “Creative and Storytelling Prompts” lesson take?

Most CoddyKit lessons take about 5–10 minutes. Each one is bite-sized and interactive, so you make steady progress and pick up exactly where you left off across the web and the app.

Can I write and run code in this AI Prompt Engineering lesson?

Yes. Every AI Prompt Engineering lesson includes a built-in code editor, so you write and run real code right in your browser and get instant AI feedback — no local setup required.

All lessons in this course

  1. Email and Professional Writing Prompts
  2. Social Media Content Prompts
  3. Technical Documentation Prompts
  4. Creative and Storytelling Prompts
← Back to AI Prompt Engineering