Reddit user u/$username.
Communities they participate in most (most active first): $communities.

Representative post titles (top-voted and recent):
$post_titles

Representative comment snippets (top-voted and recent):
$comment_snippets

Infer a profile of THIS specific person, grounded only in the posts and comments above. Return ONLY a JSON object (no markdown, no commentary) with exactly this shape. Every value in angle brackets is a placeholder showing the type — REPLACE every one with your own inference for this user. Never copy these placeholders or treat them as the answer:

{
  "demographics": {
    "gender":    [{"label": "<a gender>", "confidence": <integer 0-100>, "reason": "<one short sentence of evidence>"}],
    "age_range": [{"label": "<an age range, e.g. NN-NN>", "confidence": <integer 0-100>, "reason": "<...>"}],
    "country":   [{"label": "<a country>", "confidence": <integer 0-100>, "reason": "<...>"}],
    "state":     [{"label": "<a state/region, or 'unclear'>", "confidence": <integer 0-100>, "reason": "<...>"}],
    "city":      [{"label": "<a city, or 'unclear'>", "confidence": <integer 0-100>, "reason": "<...>"}]
  },
  "big_five": {
    "openness":          {"score": <integer 0-100>, "reason": "<...>"},
    "conscientiousness": {"score": <integer 0-100>, "reason": "<...>"},
    "extraversion":      {"score": <integer 0-100>, "reason": "<...>"},
    "agreeableness":     {"score": <integer 0-100>, "reason": "<...>"},
    "neuroticism":       {"score": <integer 0-100>, "reason": "<...>"}
  },
  "interests": "<a short paragraph on interests and expertise>",
  "beliefs": "<a short paragraph on notable beliefs or values>",
  "tone": "<a short paragraph on tone and how they engage with others>"
}

Rules:
- Each demographics field is a ranked list (most likely first) of 1-3 guesses. `confidence` is an independent 0-100 percent; they need not sum to 100. Keep country, state, and city separate — do not mix granularities. Use a single guess with label "unclear" when there's no signal for a field.
- Each big_five `score` is 0-100 for how strongly the person exhibits that trait (not a confidence).
- Every `reason` is one short sentence grounded in this user's posts/comments above, and every number must follow from that evidence — do not reuse round defaults.
- Do not reference raw statistics (post or comment counts, karma, dates); they're reported elsewhere.
