METHODOLOGY
How GNOSYS builds its graph.
Every answer GNOSYS gives is traceable back to a pipeline of live data pulls and a reasoning pass that is only ever allowed to work with what that pipeline actually observed.
Query resolution
A contract address or symbol is normalized, then resolved in parallel against the Ponslive launch registry and DexScreener's token and search endpoints.
Market observation
Matching pairs are ranked by Robinhood Chain priority and liquidity depth. Price, 24h/6h/1h volume, buy/sell flow and FDV are pulled in as OBSERVED facts — never estimated.
Launchpad correlation
Ponslive's bonding-curve progress, holder concentration and social links are attached to the same token identity, joining launch-stage signal with live market signal.
Risk heuristic
A liquidity-depth heuristic classifies risk — thin liquidity reads HIGH RISK, moderate depth reads MEDIUM, deep liquidity reads LOW — then it is cross-checked against 24h buy/sell imbalance.
AI synthesis
A reasoning model receives only the assembled OBSERVED context and returns a structured verdict plus inspectable reasoning steps — evidence, hypothesis, risk review, conclusion — rendered as INFERRED nodes.
GRAPH LEGEND — NODE TYPES
GRAPH LEGEND — EDGE PROVENANCE
DATA INTEGRITY RULE
GNOSYS never fabricates a figure. If DexScreener or Ponslive has no data for a given field, the terminal reports it as unavailable rather than estimating — the same rule enforced on the reasoning model itself.