Documentation
The Attention Engine
How LoreonLabs turns raw, live signals into a single, comparable read on what's gaining attention — computed entirely from real data.
What it does
The attention engine ingests live signals from every connected source, normalizes them into a shared model, and ranks what is accelerating. It never invents numbers — every figure traces back to a real source.
Inputs
- Market data — prices, market cap, and 24h/7d/30d change.
- Developer activity — commits, contributors, stars, and recency.
- News & social — articles across crypto and tech feeds plus developer forums.
- Web extraction — official sites, docs, and changelogs.
From signal to intelligence
- Ingest live data from each source.
- Normalize into narratives, builders, projects, ecosystems, and markets.
- Correlate entities to the themes and ecosystems they belong to.
- Rank by a concrete, real metric per surface.
- Surface the result, with provenance on every item.
Weighted toward acceleration
The engine favors momentum over size: a small project moving fast can outrank a large but flat one. The exact metric for each surface is documented in its ranking page.