About Nexus Control

Governed autonomy for on-chain finance — explained

Sign in
What is Nexus Control?

Nexus Control is an AI-agent orchestration layer for on-chain finance. It lets institutions run autonomous software agents against a portfolio of tokenised assets — while keeping every action inside a strict, human-defined governance framework.

Agents can analyse, propose and (within limits) execute portfolio actions, but they never operate unchecked: every proposal passes through a policy engine, material decisions require human approval, and everything is recorded in an immutable audit & explainability log. Think of it as the compliance and control room that sits between AI agents and your capital.

What is it used for?
Automated portfolio management

Delegate routine work — yield hunting, rebalancing, risk trimming — to agents that run on a schedule or on demand, without giving up final say on material trades.

Regulated / institutional environments

Position limits, protocol allow-lists, daily trade caps and risk-score ceilings are enforced before execution — designed for teams that answer to compliance.

Human-in-the-loop oversight

Anything above your approval threshold lands in the Action Queue for a human decision. Approve, reject with a note — every choice is logged.

Explainability & audit

Every agent run, policy check, block and portfolio change is written to a filterable audit trail with plain-English reasoning — ready for internal review or regulators.

How it works
  1. Agents propose. You run an agent (or it runs on schedule). It analyses the portfolio and generates action proposals — buy, sell, rebalance or transfer — each with a reason and a risk score.
  2. Policies filter. The policy engine checks every proposal against your rules: maximum position size, allowed protocols, daily trade limits and risk-score ceilings. Violations are blocked outright and logged.
  3. Humans decide. Proposals above your approval threshold wait in the Action Queue for sign-off. Smaller actions within policy can be auto-approved.
  4. Everything is recorded. The audit log captures who (or what) did what, when, and — crucially — why.
Governance first: the platform's core principle is that autonomy is earned inside guardrails — agents get freedom only up to the limits you set.
The four agent types
Agent typeWhat it does
Yield Optimiser Scans holdings for under-performing yield and proposes moving a portion into higher-yielding assets.
Portfolio Rebalancer Watches allocation drift against your target weights and proposes rebalancing trades when positions stray too far.
Risk Monitor Tracks portfolio-weighted volatility and proposes de-risking trades — selling the riskiest asset into the safest — when thresholds are breached.
RWA Manager Maintains your real-world-asset mandate: tops up tokenised RWA exposure from stablecoin reserves and sweeps excess cash to custody.
This is a simulation environment. All assets, prices, yields and blockchain interactions are mocked — no real funds are at risk and no real transactions occur. The platform demonstrates the governance model: agent orchestration, policy enforcement, human approval and audit trails.