Demeterics Council icon

Demeterics Council

Get diverse feedback on any content from 18 AI personas representing different US demographics

Demeterics Council

AI-generated

Summary

Evaluate given content by querying a configurable number of AI personas representing diverse US demographics to provide feedback on a specific question, then aggregate the responses with optional inclusion of raw persona feedback.

Inputs

  • Question (required) — The evaluation question to ask the council, e.g., "How would you rate this video hook?"
  • Content (required) — The text content to be evaluated (e.g., script, hook, text).
  • Number of Personas (required) — Number of AI personas to query, between 2 and 18, trading off perspective diversity against cost.
  • Persona Model — Model to use for individual persona evaluations (default: openai/gpt-oss-20b).
  • Aggregation Model — Model to synthesize the final aggregated response (default: openai/gpt-oss-120b).
  • Temperature — Controls response diversity from 0 to 1; higher values yield more varied persona feedback.
  • Max Tokens — Maximum tokens to allow per persona response (between 50 and 16000).
  • Include Raw Responses — Whether to include individual persona responses in the output (default true).

Output shape

a single JSON object per input item

The output JSON contains an id, aggregated score, summary, usage statistics, optional detailed voting stats, individual persona responses if requested, and error information if any occurred. Errors include HTTP status codes and hints when applicable. Each output item corresponds to one input item, paired by index.

Examples

Example 1: Rating a video script's engagement

Ask "How engaging is this video script?", provide the script text, query 8 AI personas with default models and temperature, and include raw persona feedback in the output.

Links

Discussion