An AI-native research and trading stack that shows how it reached a decision, tests the edge out of sample, and keeps an explicit permission boundary before live execution.
Auditable research · human-gated by default · deploy to your trust boundary
NexusTrade control planeConnected
trajectory_01JX9A
Aurora · institutional-equity-research
Complete
01
Plan research trajectory
Objective, constraints, model, and execution budget recorded
Logged
02
Query market and fundamental data
Tool inputs, source results, latency, and cost retained
Logged
03
Construct and validate candidate
Strategy changes linked to validation artifacts
Logged
04
Stage proposed order
Pending human approval · no broker submission yet
Blocked
walk-forward / anchored
Train → validate → out-of-sample
Complete
Fold 01TrainValidateOOS
Fold 02Train ↗ValidateOOS
Fold 03Train ↗ValidateOOS
Fold 04Train ↗ValidateOOS
Fold 05Train ↗ValidateOOS
5Independent folds
1×OOS promotion
SavedStudy artifact
Live order approval
Strategy · mean-reversion-us
Awaiting reviewer
AAPL
Buy · market order
Human gate
Quantity120 shares
Est. notional$25,184
BrokerageConnected
Aurora staged this proposal after validation. The universal live-order guard prevents broker submission until the approval state is explicit.
The institutional control layer
Autonomy where it helps. Accountability where it matters.
NexusTrade connects agentic research to durable evidence, validation firewalls, and explicit execution controls—so a team can inspect the path from idea to order.
01 / Auditability
Reconstruct the decision.
Trace an agent's plan, model calls, tool inputs, outputs, costs, approvals, errors, and resulting portfolio activity.
Owner-scoped agent trajectories
Durable portfolio event history
JSON and CSV export surfaces
Strategy and order provenance
02 / Permissions
Keep live execution deliberate.
Agent-originated orders are staged for review by default. A universal brokerage guard refuses live submission without an explicit approval state.
Preflight before approval
Approve or reject with a recorded decision
Atomic review for grouped orders
Brokerage-independent enforcement
Control boundary: per-strategy automatic approval is available. Institutions requiring manual review keep it disabled; NexusTrade does not imply a human reviewed an order when that setting is enabled.
03 / Research infrastructure
Test the edge out of sample.
Run anchored or rolling walk-forward studies in the Rust engine with a strict train, validation, and out-of-sample sequence.
Two to eight fold studies
Genetic or systematic optimization
One-time OOS promotion
Progress, cancel, and checkpoint support
Evidence, not a black box
A review surface for every critical boundary.
The page is not claiming a compliance certification. It shows the technical records and permission boundaries a compliance program can evaluate and incorporate.
trajectory_01JX9A
Agent work can be inspected step by step.
Each trace connects agent state and plan to tool execution, model and usage metadata, approvals or rejections, errors, and run metrics. Access is scoped to the owning user.
Portfolio events link timestamps, users, portfolios, strategies, order status, assets, side, quantity, price, fees, and payload. Recent events are queried from the hot store and compacted portfolio history is available from the durable event lake.
eventType
ORDER_APPROVED
eventTimestamp
2026-08-04T14:32:08.221Z
strategyIds
["mean-reversion-us"]
export
JSON · CSV
Walk-forward v2
Candidate selection stays outside the final test.
Optimization happens on training windows, candidates are selected on validation windows, and the winner is promoted once to each out-of-sample fold. Aggregate results and fold artifacts persist for review.
An agent proposal begins unapproved. Preflight exposes the order for review, and an explicit decision is recorded before live submission. The final brokerage adapter rechecks approval rather than trusting an upstream UI.
approvedByUser
false → true only after approval
default status
Pending User Approval
enforcement
Universal live-order brokerage guard
automation
Optional per strategy; explicit and reviewable
Choose the operating model
Fit NexusTrade to your trust boundary.
Start with an embedded agent, run the platform inside infrastructure you control, or connect proprietary data and execution systems.
Embed the experience
Whitelabel Aurora
Bring the research and strategy-building agent into your own product, brand, and customer journey.
Embedded agent experience
Scoped partner sessions
Brand and launch support
Own the deployment
Private NexusTrade
License and operate the NexusTrade stack on infrastructure you control so proprietary strategies stay inside your environment.
Customer-controlled infrastructure
Private strategy boundary
Deployment and engineering support
Connect your stack
Data & execution integration
Integrate proprietary datasets, internal research systems, brokerage connections, and team-specific workflows.
Market and alternative data
Brokerage and execution adapters
Workflow and platform integration
Engagement model
From trust boundary to production path.
Institutional work begins by mapping the real decision, data, and execution boundaries—not by forcing a generic deployment.
Architecture review
Map users, agent permissions, data ownership, brokerage connections, and required review surfaces.
Proof of integration
Validate the highest-risk data, identity, or execution boundary against a narrow workflow.
Controlled rollout
Ship with scoped sessions, approval defaults, telemetry, and an agreed operating model.
Evidence review
Review agent traces, validation artifacts, portfolio events, and approval behavior with your team.
Build against the real boundary
Give your team a system it can inspect.
Tell us where NexusTrade needs to run, what it needs to connect to, and who must approve live execution.
NexusTrade provides technical controls that may support an institution's compliance program. It does not represent regulatory approval, compliance certification, legal advice, or investment advice. Institutions remain responsible for configuring and supervising their own policies, permissions, and execution workflows.
Institutional inquiry
Tell us what you are building.
We will use this to scope the right deployment and follow up directly.