Skip to content
Signals
deepseek-ai/deepseek-harness adds ~58,495 stars this week — breakout in TypeScriptTrendNexus signalsstorytold/pdfcraft adds ~27,763 stars this week — emerging in RustTrendNexus signalsEbony-Vinyl/dsh-our-free-model adds ~19,960 stars this week — emerging in JavaScriptTrendNexus signalseternity4719/HowToLiveBetter adds ~19,906 stars this week — breakout in HTMLTrendNexus signalsgarrytan/gstack adds ~15,261 stars this week — breakout in TypeScriptTrendNexus signalsGraphify-Labs/graphify adds ~15,142 stars this week — breakout in PythonTrendNexus signalsodysseus-dev/odysseus adds ~12,438 stars this week — breakout in PythonTrendNexus signalsPanniantong/Agent-Reach adds ~12,219 stars this week — breakout in PythonTrendNexus signalsRisk watch: bagisto/bagisto trends cooling with security score 92TrendNexus signalsRisk watch: thedaviddias/Front-End-Checklist trends cooling with security score 82TrendNexus signalsHealthiest in cohort: ultraworkers/claw-code holds PVI 84 — sustainability leaderTrendNexus signals

Nexus metric

Nexus Velocity™

Acceleration against the subject's own baseline — not absolute size.

4 reading bandslive across 12 tracked signals

What it measures

Nexus Velocity expresses how fast a subject's signal volume is growing relative to its own history, as a multiplier. 2.0× means twice the subject's normal signal rate; 1.0× means business as usual. Because it is self-referential, a small subject accelerating hard outranks a large subject coasting.

Absolute counts favor incumbents — big subjects always have more posts, stars and coverage. Velocity is the size-blind lens that surfaces the next thing while it is still small.

How it's computed

Signal rate in the recent window divided by the trailing baseline rate for the same subject, with a floor to avoid divide-by-zero on new entities. Displayed as a rounded multiplier; inside Nexus Signal it enters log-scaled at 20% weight — the heaviest single factor.

Recent signal ratecomponent

Signals per day across all ecosystems in the recent window.

Every qualifying signal counts once, regardless of ecosystem.

Trailing baselinecomponent

The subject's own normal, from the prior period.

Signals per day in the trailing window preceding the recent one.

How to read it

3.0×+

Surging

Exceptional acceleration — usually accompanies an event: a launch, a paper, a viral thread.

1.5–3.0×

Accelerating

Clearly above its own baseline. The classic early-breakout profile.

0.8–1.5×

Steady

Moving at its normal rate.

< 0.8×

Cooling

Below its own baseline — attention receding.

Velocity is noisy for very new subjects with tiny baselines; Nexus Confidence discounts exactly that case. Read a huge multiplier on thin evidence as a prompt to look, not a conclusion.