Nova AI
bySee the market before acting on it.
Tracks spot and derivatives markets alongside on-chain flows. Each signal keeps the data window, model version and backtest context that produced it.
Research output · no autonomous execution
Nova watches the market, Atlas prepares the decision and Aegis checks the order before it reaches a venue.
Use one system on its own or connect the full path. The evidence and risk state travel with the decision.
Spot, derivatives, order books and on-chain events arrive with timestamps and health checks.
Finds market regimes and signals, with the source window and backtest kept attached.
Builds a cited brief around portfolio context, uncertainty and the team’s limits.
Checks exposure, liquidity and venue state. A failed rule stops the order.
Only an approved order reaches the venue. The decision and operator remain traceable.
Each system can operate independently. Together they connect evidence, interpretation and capital protection.
Tracks spot and derivatives markets alongside on-chain flows. Each signal keeps the data window, model version and backtest context that produced it.
Research output · no autonomous execution
Turns signals into research briefs for a desk or client team. It cites the inputs, shows uncertainty and never places an order.
Decision support · never submits an order
Checks proposed orders against exposure, liquidity, venue health and hard limits. If a feed is stale or a rule breaks, execution stops and the reason is logged.
Policy enforcement · blocks on stale state
Ava, Elara and Iris provide cited knowledge, controlled automation and release evaluation across the product line.
A permission-aware knowledge workspace that turns approved documents into cited answers and versioned context for people and AI tools.
eval-2026-04-17 passed the access-control rerun.
Evaluation plan §4.2Review record #184A bounded operations assistant that classifies requests, gathers approved context and prepares the next action for a human reviewer.
An evaluation workbench for comparing prompt, model, retrieval and tool changes against a team’s own versioned test cases.