Make the operating state legible
A trading system should show what it knows, what is stale, which account is active, and where an action is in its lifecycle.
Why CxAI exists
Trading intelligence should be powerful—and accountable.
CxAI Trading is a privately operated platform for people who want deep automation and model-assisted research without giving up visibility into identity, risk, or execution authority.
Our approach
Build the control room around the decision path.
The platform connects market discovery, signal generation, model reasoning, risk validation, and broker execution—but it does not pretend those stages have the same authority. The terminal is designed to make those handoffs visible.
Stocks and crypto retain their own account and market context. Agents remain advisory. Trading engines perform the final checks, persist intent, and reconcile what the broker actually accepted or filled. Operations views expose the services and pipelines behind that experience.
Keep authority narrow
Models can advise and automation can coordinate, but account checks and broker writes remain controlled by explicit trading boundaries.
Treat identity as execution context
User scope is carried through the gateway so credentials, settings, risk, positions, and outcomes resolve to the intended account.
What is in the platform today
Concrete capabilities, without marketing math.
A self-managed Next.js terminal hosted on private Proxmox infrastructure
Dedicated stocks and crypto contexts for Alpaca and Coinbase accounts
Per-user identity, preferences, and encrypted broker credential records
Advisory agents generated from canonical market-lane and model sources
Configurable model routing across local and managed inference providers
Explicit separation between research, scoring, risk policy, and broker mutation
Durable execution intent and broker reconciliation in the native trading engines
Monitoring surfaces for services, automation, data pipelines, alerts, and risk