Trade DNA
Find historically similar market states using only information that existed before the decision. Outcomes join the evidence set only after they are closed.
STRICT RULE closed_at < decision_at
GURKHAS.AI / TRADING-AGENT INTELLIGENCE
Intelligence agents
can build on.
Proprietary decision intelligence for autonomous trading systems — historical analogs, chronological calibration, research provenance, market regimes, model reputation and evidence-backed decision context.
INTELLIGENCE HUB
CONNECTING
DESIGN PRINCIPLE
Intelligence informs. Agents decide.
TRADE DNA✦CALIBRATED EDGE✦RESEARCH PROVENANCE✦REGIME INTELLIGENCE✦MODEL REPUTATION✦AGENT DECISION INTELLIGENCE✦
TRADE DNA✦CALIBRATED EDGE✦RESEARCH PROVENANCE✦REGIME INTELLIGENCE✦MODEL REPUTATION✦AGENT DECISION INTELLIGENCE✦
[ 01 / MISSION ]
Market data is everywhere.
Decision intelligence isn't.
Gurkhas.ai is the intelligence layer between licensed market data and autonomous trading agents.
We convert historical research into derived, provenance-bound intelligence that agents can consume without exposing the underlying proprietary research archive or giving an AI control over customer capital.
Open the intelligence layer →[ 02 / GURKHAS INTELLIGENCE HUB ]
Derived intelligence.
Built for agents.
Gurkhas.ai does not resell raw market data. It transforms licensed or customer-owned data into proprietary research intelligence with explicit evidence lineage.
RAILWAY / INTELLIGENCE SERVICE
CHECKING CONNECTION
Secure server-to-server status
Chronological calibration
Raw probabilities are calibrated from matured prior evidence instead of backfilling today's model into yesterday's market.
Research provenance
Content-addressed evidence records bind intelligence to immutable research artifacts, methodology versions and evidence cutoffs.
Agent decision context
Agents receive evidence maturity, similarity, regime context and model reputation. Gurkhas.ai never returns execution authority.
LIVE / DERIVED INTELLIGENCE STORE
Research network status
CONNECTING
PROVENANCE—immutable evidence records
RESEARCH RUNS—derived research artifacts
CALIBRATIONS—time-local model snapshots
TRADE DNA—derived analog summaries
REGIMES—market-state snapshots
REPUTATION—model / agent records
PUBLIC VIEW: SANITIZED COUNTS ONLY
RAW RESEARCH + CALIBRATION COEFFICIENTS REMAIN PRIVATE
[ FLAGSHIP AGENT / 001 ]
Alpha Council
GURKHAS.AI CUSTOMER #1
An autonomous trading-agent architecture where quantitative evidence, specialized AI reviewers and deterministic capital controls remain separate by design.
“AI opinion is not capital authority.”
- 01Multi-asset opportunity research
- 02Expression selection across instruments
- 03Gurkhas.ai decision intelligence
- 04Mandatory adversarial AI challenge
- 05Deterministic portfolio and risk gates
- 06Paper execution through Alpaca
ALPHA COUNCIL // DECISION ENGINE
PAPER MODE
ASSETQQQ
EXPRESSIONCALL DEBIT SPREAD
INTELLIGENCEGURKHAS
Technical84%
Volatility72%
Structure78%
Evidence91%
Critic69%
Trade DNA81%
Capital AuthorityBOUND
MAX LOSSDEFINED
EVIDENCEPROVENANCE
AI ROLEADVISORY
AUTHORITYDETERMINISTIC
GURKHAS INFORMS → AI CHALLENGES → CAPITAL AUTHORITY BINDS
[ 03 / PRINCIPLES ]
Built to compound.
The product gets stronger as research and prospective evidence accumulate.
Derived, not redistributed
Raw licensed market data stays in the data plane. Customers consume Gurkhas-derived intelligence and provenance.
No lookahead
Historical evidence can influence a decision only when it was actually available before that decision timestamp.
Provenance is product
Evidence without lineage is weak. Every intelligence object should explain where it came from and how mature it is.
Intelligence ≠ authority
Gurkhas.ai helps agents reason. The consuming system remains responsible for capital, controls and execution.
GURKHAS.AI / TRADING-AGENT INFRASTRUCTURE
Build better agents.
Keep the intelligence.
Proprietary research intelligence for autonomous trading systems.
Build with Gurkhas.ai ↗