Where they are
Personas in the sidebar (/personas), and inline on the Simulate page.
Writing one
A persona is free text — brief it like an actor. The generated ones give a sense of the useful shape:
- Occupation and situation
- Demographics — age, location, household
- Goals and needs — what they wanted from the product
- Constraints and frictions — what got in the way
- Voice — how they talk, how forthcoming they are
Detail is what makes a simulation worth reading. A vague persona produces bland, agreeable answers.
Generating
Generate drafts personas from the study's own context: its type, description, research question, objectives and target population. Choose how many. They are marked AI Generated, and you can edit them afterwards.
Because generation reads the study details, good study details produce better personas.
Managing
Edit or delete any persona. Deleting one keeps interviews already simulated with it — and editing one does not rewrite past runs, which keep the text they ran against.
With MCP
list_personas {"study_id": "…"}
create_persona {"study_id": "…", "content": "…", "source": "…"}
update_persona {"study_id": "…", "persona_id": "…", "content": "…"}
delete_persona {"study_id": "…", "persona_id": "…"}
generate_personas{"study_id": "…", "count": 3, "save": true}
generate_personas with save: false returns drafts without storing them, which is useful when you want to review or edit before committing.
Last updated: September 2026