KOANO ingests data from dozens of sources — free public datasets live today, plus commercial providers that come online as they are licensed. Every verdict cites its top data sources. We publish the full list here because trust requires transparency.
25+
Free sources
20+
Paid sources
24h
Max data age
100%
Source-cited verdicts
These sources are ingested on a rolling schedule via GitHub Actions cron jobs and stored in Supabase. They form the backbone of every verdict.
These sources are called when a verdict is requested and cached for 24 hours. Their data is the signal layer that makes the reasoning accurate at the neighborhood level.
Free sources are ingested via GitHub Actions cron jobs running 24/7. Paid sources are called on-demand and cached for 24 hours to minimize API costs.
Every signal is normalized to a common schema: location (lat/lng + census tract), timestamp, data type, confidence level, and source attribution.
Normalized signals are stored in Supabase with Row Level Security enforced. Cluster 5 data lives in an isolated schema per enterprise client — never co-mingled.
When a verdict is requested, relevant signals are assembled and passed to the five specialist agents simultaneously. Each agent reasons independently.
The synthesis agent receives all five structured outputs, arbitrates conflicts, and issues the final verdict with a full reasoning chain.
Every KOANO verdict cites its top data sources. Request early access to see how dozens of sources come together in a single verdict.