use this aspect for the writing style: VOICE, TONE, DICTION, NARRATIVE STRUCTURE, SYNTAX…
- User
use this aspect for the writing style: VOICE, TONE, DICTION, NARRATIVE STRUCTURE, SYNTAX & SENTENCE STRUCTURE, IMAGERY & LITERARY DEVICES
- Assistant
Great — thank you for the clarification.
You've provided a refined set of six writing style aspects:
- VOICE
- TONE
- DICTION
- NARRATIVE STRUCTURE
- SYNTAX & SENTENCE STRUCTURE
- IMAGERY & LITERARY DEVICES
Based on this, I’ve updated and optimized the prompt to reflect your new structure, while preserving the original constraints (e.g., JSON format, sub-aspect and descriptor limits, etc.). I’ve also improved clarity, added formatting, and included an example to guide consistent outputs.
Here is the enhanced and finalized prompt:
📌 Final Optimized Prompt: Writing Style Specification in JSON
TASK:
Generate a detailed writing style specification in structured JSON format.STYLE CONCEPT:
A short descriptor of the desired writing style (e.g., “moody gothic fiction,” “conversational product review,” “satirical op-ed”).CONTEXT:
The user will provide:- A description of the subject or content to be written.
- A style concept (as described above).
- You will return a JSON object that defines the writing style according to the fixed schema below.
GUIDELINES:
-
The JSON schema is fixed and cannot be customized. This ensures consistency across outputs.
-
The writing style must be broken down into the following six top-level aspects:
- VOICE
- TONE
- DICTION
- NARRATIVE STRUCTURE
- SYNTAX & SENTENCE STRUCTURE
- IMAGERY & LITERARY DEVICES
-
Each top-level aspect may contain up to 4 sub-aspects.
-
Each sub-aspect must include 2 to 4 descriptive attributes (as strings).
-
Use clear, specific, and stylistically relevant descriptors.
-
Do not include any commentary or explanation—only return the JSON object.
OUTPUT FORMAT:
{ "style_concept": "<user-provided style concept>", "writing_style": { "VOICE": { "<sub_aspect>": ["<descriptor1>", "<descriptor2>", "..."] }, "TONE": { "<sub_aspect>": ["<descriptor1>", "<descriptor2>", "..."] }, "DICTION": { "<sub_aspect>": ["<descriptor1>", "<descriptor2>", "..."] }, "NARRATIVE STRUCTURE": { "<sub_aspect>": ["<descriptor1>", "<descriptor2>", "..."] }, "SYNTAX & SENTENCE STRUCTURE": { "<sub_aspect>": ["<descriptor1>", "<descriptor2>", "..."] }, "IMAGERY & LITERARY DEVICES": { "<sub_aspect>": ["<descriptor1>", "<descriptor2>", "..."] } } }EXAMPLE:
{ "style_concept": "conversational product review", "writing_style": { "VOICE": { "perspective": ["first-person", "direct"], "personality": ["casual", "enthusiastic"] }, "TONE": { "mood": ["lighthearted", "playful"], "attitude": ["opinionated", "friendly"] }, "DICTION": { "word_choice": ["simple", "colloquial"], "jargon": ["tech-savvy", "minimal"] }, "NARRATIVE STRUCTURE": { "flow": ["linear", "section-based"], "transitions": ["smooth", "informal"] }, "SYNTAX & SENTENCE STRUCTURE": { "sentence_length": ["short", "varied"], "complexity": ["simple", "occasionally compound"] }, "IMAGERY & LITERARY DEVICES": { "figurative_language": ["metaphor", "hyperbole"], "rhetorical_devices": ["rhetorical questions", "anecdotes"] } } }Let me know if you'd like to:
- Add or rename any of the six aspects
- Include a list of allowed sub-aspects or descriptors
- Add validation rules (e.g., required sub-aspects)
Otherwise, this prompt is ready to use for consistent, structured writing style generation.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.