Skip to content

Trend Finder

Trend Finder is a repo-local AI OS extension for collecting reviewed public signals, normalizing evidence, ranking topics, and turning those findings into creator decisions.

Its goal is not merely to report popularity. It looks for topics that are early enough, evidenced enough, and relevant enough to the current Creator Lens to act on or monitor.

  1. Inspect source health and freshness.
  2. Review topics, evidence, score breakdowns, and provenance.
  3. Use Creator Lens and hidden-gem views to find differentiated angles.
  4. Track promising topics in the Watchlist.
  5. Export or share a bounded Brief.
  6. Use Engine Replay to inspect sanitized latest-run stages.

Workbench adds search, comparison, triage, notes, tags, and next-action controls. Engine Replay shows sanitized latest-run stages; it is not a historical backtest or provider trace viewer.

Each source needs a reviewed compliance posture. Collection should preserve provenance, distinguish source items from normalized evidence, and make score components inspectable. A high score is a decision aid, not a factual guarantee.

Direct public adapters, reviewed configured sources, quality caps, health states, deterministic fallback, and AI analysis are labeled independently.

The local extension can run configured collectors and scoring workflows. The static demo contains reviewed Trend Finder fixtures only: it does not host collectors, analysis APIs, or write paths.