Hacker News momentum tracker and theme clustering dashboard
Build a workflow called Signal Oracle that turns the Hacker News front page into an overnight intelligence briefing — ranking stories by momentum (how fast their score is accelerating) rather than raw score, and clustering the page into the themes it's collectively talking about. The goal is to surface what's genuinely gaining traction, not just what's already popular. Trigger and structure: On an hourly cron schedule, fetch the Hacker News front page from a public source (I used the Algolia HN API — or any HN data source you prefer) and append a timestamped snapshot to a persistent volume that accumulates history (cap it at a few days of snapshots). Momentum needs history, so this step builds it over time. Expose a JSON API route (e.g. GET /signal) that reads the snapshot history, computes per-story score velocity, a blended signal score, and theme clusters, and returns the report. On the first run when there's only one snapshot, fall back gracefully to a freshness-weighted heat score and let the ranking sharpen as history builds. Expose a webpage route (e.g. GET /oracle) that fetches the report and renders it as a dashboard — a momentum leaderboard plus a theme constellation/visualization. Make it a polished, distinctive page. Keep the HTTP routes authenticated to the space. No credentials are needed since the data source is a public, no-auth API — the only writes are snapshots to the workflow's own volume; everything against HN is read-only. Tools used: Algolia Hacker News API
What this prompt builds
Signal Oracle transforms the Hacker News front page into a momentum-based intelligence briefing by tracking score velocity over time rather than raw popularity. An hourly cron job snapshots the front page, computes acceleration for each story, and clusters content into dominant themes. Founders, engineers, investors, and product teams get a dashboard showing what's rising now — not what already peaked — letting them spot trends earlier with a glance instead of manual scanning.
The problem
The Hacker News front page ranks by raw score, which heavily biases toward stories that have already peaked — by the time something reaches the top, the moment has often passed. Founders, engineers, investors, and product teams who follow HN to stay ahead of the curve end up manually refreshing and eyeballing the page, yet still catch trends late. There's no easy way to see what's accelerating right now or to step back and grasp the themes the page is collectively discussing. Signal Oracle solves this by computing score velocity from hourly snapshots, blending momentum into a signal score, and clustering stories into dominant themes — delivering a dashboard that surfaces what's genuinely gaining traction, not just what's already popular, so users spot rising stories earlier and read the day's themes at a glance.
Solution and impact
Signal Oracle turns the HN front page into an overnight intelligence briefing. An hourly cron job snapshots the page into a history volume, then a signal engine computes each story's score velocity and clusters the page into its dominant themes, rendering a momentum leaderboard and theme constellation in the browser. Instead of manually scanning HN and catching trends after they've peaked, users get a ranked view of what's actually gaining steam plus the day's themes at a glance — a few seconds instead of a scroll-and-guess routine, spotting rising stories earlier. The workflow needs no credentials, runs on a public read-only API, and sharpens its ranking automatically as hourly snapshots accumulate.