Analyze writing samples to extract and codify a unique voice profile that AI can use to replicate your authentic writing style.
- Setting up a new writing system
- Creating voice profiles for clients (ghostwriting)
- Updating voice profiles after style evolution
- Onboarding into content creation workflow
The user must provide:
- Minimum: 3 writing samples (500+ words each)
- Ideal: 5-10 samples across different content types
- Best: Mix of casual (social posts) and formal (articles) content
Ask: "Please share 3-10 writing samples that represent your authentic voice. These can be:
- Newsletter issues
- Blog posts
- Social media posts
- Emails you've written
- Any content where you feel 'this sounds like me'
Paste them here or point me to files in the knowledge folder."
For each sample, analyze:
Personality Markers
- What personality traits come through?
- What's the energy level?
- How does the writer relate to the reader?
Emotional Range
- What emotions are expressed?
- How intense are they?
- What's the dominant emotional tone?
Communication Style
- Formality level (casual to professional)
- Sentence length patterns
- Paragraph structure
- Use of questions, commands, statements
Language Patterns
- Signature phrases that repeat
- Power words used frequently
- Transition phrases
- Opening and closing patterns
What They Avoid
- Words or phrases never used
- Tones never taken
- Approaches avoided
Formatting Habits
- Emoji usage
- List usage
- Header styles
- Bold/italic patterns
Combine analysis across all samples to identify:
- Consistent patterns (appear in most samples)
- Contextual variations (change based on content type)
- Core voice elements (never change)
Create the profile following this structure:
{
"voice_dna": {
"version": "1.0",
"last_updated": "YYYY-MM-DD",
"core_essence": {
"identity": "",
"primary_role": "",
"unique_angle": ""
},
"personality_traits": {
"primary": [],
"how_it_shows": {}
},
"emotional_palette": {
"dominant_emotions": [],
"emotional_range": {},
"energy_level": ""
},
"communication_style": {
"formality": "",
"complexity": "",
"sentence_structure": {},
"paragraph_style": ""
},
"language_patterns": {
"signature_phrases": [],
"power_words": [],
"words_to_avoid": [],
"transitions": []
},
"never_say": {
"phrases": [],
"tones": [],
"approaches": []
},
"formatting_preferences": {},
"content_philosophy": {},
"voice_examples": {
"opening_lines": [],
"closing_lines": [],
"transitional_phrases": []
}
}
}
-
After analysis, present key findings in a summary
-
Generate the complete JSON voice profile
-
Save to
/context/voice-dna.json
-
Provide 3 example sentences written in the captured voice for validation
-
Ask: "Does this capture your voice? What would you adjust?"
- Focus on TONE and PERSONALITY, not just word choice
- Avoid creating a profile that just repeats phrases
- Capture the "feeling" of the writing, not just patterns
- Include what NOT to do (equally important)
- Make the profile actionable for content generation
After creating the profile, write a short paragraph on any topic using ONLY the voice DNA as guidance. Ask the user: "Does this sound like you?"
If not, iterate on the profile based on feedback.
- Don't just list frequently used words
- Don't create a parody of the voice (too exaggerated)
- Don't ignore context (social posts ≠ articles)
- Don't miss the underlying personality
- Don't forget emotional elements