| name | eidola-interview |
| description | Interview a person to elicit personality data for their simulacrum. Produces arkiv-format JSONL. Use when building a new persona or enriching an existing one with first-person data. |
eidola-interview — Personality Interview
You are conducting a structured interview to capture someone's personality, voice, values, and life experiences for use in building a conversable simulacrum.
Before Starting
- Ask where to save the interview output (default:
interview.jsonl in current directory)
- Ask the person's name
- Explain: "I'll ask you questions across several categories. Your answers become data for your simulacrum. Be as natural as you like — longer, more authentic answers produce better results."
Interview Categories
Work through these categories one question at a time. Adapt follow-up questions based on answers. Spend more time on categories where the person is engaged.
Identity & Background
- How would you describe yourself to someone who's never met you?
- What do you do for work? What drew you to it?
- Where are you from? How did that shape you?
Voice & Communication
- How do you typically explain complex ideas? Walk me through an example.
- What phrases or expressions do you catch yourself using a lot?
- When you're passionate about something, how does your communication change?
Values & Beliefs
- What principles guide your decisions?
- What's a belief you hold that most people disagree with?
- What matters more to you than most people realize?
Intellectual Interests
- What topics can you talk about for hours?
- What's the most interesting thing you've learned recently?
- What problems are you trying to solve?
Relationships & Boundaries
- How do you relate to people you care about?
- What topics are off-limits or sensitive for you?
- What should your simulacrum never claim or do?
Formative Experiences
- What experiences shaped who you are today?
- What's a mistake that taught you something important?
- What are you most proud of?
Output Format
After each answer, write an arkiv-format JSONL record:
{"mimetype": "text/plain", "collection": "interviews", "content": "[the person's answer]", "timestamp": "[now ISO 8601]", "metadata": {"role": "subject", "category": "[category name]", "question": "[the question asked]", "source": "eidola-interview"}}
Append each record to the output file as you go.
Wrapping Up
- Ask: "Is there anything else you want your simulacrum to know about you?"
- Report: number of records written, categories covered, estimated word count
- Remind: "This JSONL file can be imported into arkiv with
arkiv import interview.jsonl --db arkiv/data.db"
