OMNIS
Daily breakout signal detection

Find breakout startups before consensus does.

Omnis continuously scores early-stage companies across GitHub, npm, Hacker News, Product Hunt, job boards and macro regime data — surfacing the next breakout while it's still mispriced.

5-factor quant modelWalk-forward backtestedDaily re-scoring
universe_ranking.live live
01S███████devtools+7.6σD10
02B███████████████ I███healthtech+6.1σD10
03H█████climate+4.8σD10
04G████████ P██████████████healthtech+4.7σD10
company names hidden — sign in to reveal
The pipeline

From raw signal to ranked conviction

A closed loop that ingests public signals, scores them against a cross-sectional model, ranks the universe, and tunes itself on realized outcomes.

01
Ingest
Signal agents pull GitHub, npm, HN, Product Hunt, jobs & macro data.
02
Score
Five-factor model, z-scored within sector cohorts into deciles.
03
Rank
Cross-sectional universe ranking with managed watchlists.
04
Learn
Hit-rate feedback decays and re-tunes factor weights.
The model

Five factors, one conviction score

Momentum
Stars / forks delta, npm growth, HN velocity.
Growth
Hiring velocity, contributor growth, sustained trajectory.
Quality
Signal consistency, low noise, durable trends.
Attention
HN mentions, social growth, Product Hunt votes.
Risk
Volatility, spike-revert patterns, macro regime drag.
Composite
Blended decile 1–10, re-scored every day.
Agent-native

Built for humans and autonomous agents

Every signal is reachable over a remote MCP server and a machine-readable API. Agents can discover, pay, and pull ranked deals with no human in the loop.

agent.sh
# one paid read — ranked discovery candidates
curl https://bamboosnow.co/api/v1/discovery/top?limit=10 \
  -H "Authorization: Bearer $OMNIS_KEY"

# → scored, decile-ranked breakout candidates
{ "rank":1, "company":"Vellum Labs", "decile":9, "z":18.4 }

See what's breaking out today

Daily, decile-ranked conviction on the companies that matter — before they're consensus.