ARX One Platform

The control plane for multi-agent enterprise.

ARX One puts one governed layer in front of every model and agent you run. Route them, secure them, account for the spend, and carry memory across all of them, from a single platform.

Drop-in with the OpenAI and Anthropic APIs you already use. Built for finance, healthcare, and legal.

Backed by

Product & features

One product. Every model and agent. One control plane.

ARX One is the whole platform. Everything below is a feature of it, governed by the same control plane. Not six tools to integrate, and not six separate purchases.

ARX One

The unified control plane for every model and agent in the enterprise. Run it, govern it, and account for it from one place, with role-based access and full multi-tenancy built in.

Unified control planeCentralized visibilityRole-based accessEnterprise multi-tenancy
Routing

Model & agent routing

One policy-aware connection to every model and agent, with intelligent routing, request validation, and telemetry built in.

Intelligent routingRequest validationGuardrailsObservability
FinOps

Cost visibility & control

Price and attribute AI spend to a team, forecast it, and optimize cost without losing quality.

Cost analyticsChargebackBudgets and forecastingOptimization
Governance

Policy & access control

Policy, risk, and access guardrails enforced continuously on every request, not checked after the fact.

Policy managementAccess governanceRisk controlsCompliance automation
Evidence

Audit-ready evidence

Immutable lineage for every decision, mapped to the frameworks you report against and ready for audit.

Lineage trackingImmutable recordsAudit trailTamper-proof storage
Memory

Cross-model memory

Memory that survives model upgrades and compounds over time, so context and knowledge move with the work.

Cross-model memoryContext portabilityTeam RBACVerified state transfer
Marketplace

Model & agent marketplace

A governed catalog of trusted models and agents, from selection through licensing and secure acquisition.

CatalogTrust and ratingsLicensingSecure acquisition

Connect anything, govern everything

One layer in front of your whole AI estate.

Models, agents, data, and the people who use them, all connected through ARX and governed by the same rules.

Users and systems

Your people and applications, with single sign-on and role-based access.

AI models

Hosted, bring-your-own, and marketplace models through one interface.

Agents

Internal, third-party, and custom agents, coordinated and governed.

Data and tools

Enterprise data, APIs, and retrieval sources, connected safely.

Infrastructure

Cloud, on-prem, hybrid, or sovereign. You choose where it runs.

External ecosystem

Partners, vendors, and regulators, with the evidence they ask for.

How it works

Control on day one. Keep your models.

ARX speaks the OpenAI and Anthropic wire formats. Point your app at the platform and control starts on the first call.

1

Connect

Change one line. Send traffic through ARX instead of straight to the provider. No SDK rewrite.

2

Govern and meter

Policy, budgets, and access run in the path. Every call is priced and tied to a team.

3

Prove and remember

Every decision is evidenced and audit-ready, and memory builds from the first call.

# one line
client = OpenAI(base_url="https://api.arxone.ai/v1")

Get started

Bring every model and agent under one control plane.

See what you spend, govern what your models and agents do, and keep what your teams learn. We will map it to the regulation you answer to and walk you through the first call.