Enterprise Early Access — Limited Cohort

The Enterprise
Agent Forge

NOVA is roaiq's production platform for deploying governed AI agent workforces at enterprise scale — multi-tenant, auditable, and built for the governance requirements your legal team will actually approve.

Sign up for NOVA → Explore capabilities
Multi-tenant isolation
Full execution audit trail
Matimo governance built-in
16 reasoning strategies — selectable per agent
NOVA / Workflow Studio — Customer Onboarding Agent v2.1 LIVE EXECUTION
Node Library
⚡Trigger
⇄Router
◈LLM Engine
⚙Tool Call
↗Output
Governance
🛡 Matimo Policy
📋 Audit Log
⚡
Webhook Trigger
POST /v1/agent/onboard · Tenant: acme-corp
DONE
◈
Intent Classifier — GPT-4o
Engine: reasoning-v3 · Tokens: 312 / 4096
RUNNING
⚙
CRM Enrichment Tool
HubSpot API · Governed by: policy/crm-read-only
QUEUED
↗
Structured Output + Notify
Slack + Email · Schema: OnboardingResult v2
QUEUED
Execution Trace
StatusRunning
Latency1.24s
Tokens used312
Cost$0.0014
Tenantacme-corp
AuditLogged ✓
Governance
🛡 Matimo Active policy/crm-read-only
policy/pii-redact-v2
2 policies enforced
The problem space

Enterprise AI deployment
is broken at the seams

Most AI orchestration tools are built for demos or developer experiments. When enterprises try to productionise them, three walls appear immediately.

Governance is bolted on, not built in

Policy enforcement, tool access controls, and audit trails are afterthoughts — left to teams to wire together manually after the system is already running.

Multi-tenancy breaks at scale

Self-hosted solutions require every tenant to be a separate deployment. Data leaks between contexts. Operational overhead compounds fast.

Infrastructure costs are unpredictable

Always-on agent infrastructure costs money even when nothing is running. Per-seat licensing punishes growth. There's no model aligned with actual usage.

Platform capabilities

Everything enterprise deployment
actually requires

NOVA is not a framework. It's a production system — designed for the constraints that enterprises actually operate under.

16 Reasoning Engines — Choose the Right Strategy

NOVA supports 16 distinct reasoning strategies — from ReAct and Chain-of-Thought to Tree-of-Thoughts, Plan-and-Execute, Reflexion, and Multi-Agent coordination. Select the right engine for each agent based on your task profile: structured reasoning, iterative refinement, goal decomposition, or multi-step planning.

OpenAI Anthropic Google Ollama Multi-LLM routing

Visual Workflow Studio

Operations teams compose multi-step agent workflows on a drag-and-drop canvas. Engineers extend and integrate via REST API. Both interact with the same execution engine — no duplication of logic.

Drag-and-drop canvas REST API

Full Execution Audit Trail

Every token, every tool call, every decision — logged with tenant context and policy state. Exportable for SIEM integration. Built to support compliance requirements as they mature.

Per-tenant logging Exportable logs

Native Multi-Tenant Isolation

Tenant context is enforced at the execution layer — not as application logic. Data, tool access, and execution budgets are isolated by default.

Tenant scoping Context isolation

Human-in-the-Loop Approval

Agents pause on high-risk tool calls and wait for a human to approve or reject before proceeding. The approval queue is built into the admin dashboard — no deployment required to gate any action.

HITL gate Matimo policy engine

Tool Providers + MCP Extensibility

8 built-in Matimo tool providers — Slack, Gmail, GitHub, Notion, HubSpot, PostgreSQL, Mailchimp, and Twilio — with 248+ individual tool actions, each governed by Matimo's policy engine. Connect any MCP server for unlimited extensibility.

248+ tool actions MCP compatible

Stateful Agent Memory

Agents maintain context across multi-step workflows and long-running processes. Memory is tenant-scoped, TTL-managed, and governed — not a global mutable store.

Long-running Tenant-scoped

Operations Control Plane

A real-time admin dashboard for your AI operations team. Monitor live executions, trace failures, review usage costs by workflow and tenant, and manage HITL approval queues — all without touching the backend. The visibility your ops team needs once agents are running in production.

Live monitoring Failure tracing Cost attribution HITL approvals
Built on Matimo

Governance designed in
from the ground up.

NOVA is architected around Matimo, roaiq's open-source agent governance layer. The platform is designed so that policy enforcement, tool access controls, and audit trails are structural — not bolted on after deployment.

For enterprises that need centralised policy management, SSO/SAML, and an exportable audit registry, Matimo Enterprise is being built as an add-on subscription — designed to extend NOVA's governance capabilities to meet the full demands of regulated and security-conscious organisations.

Explore Matimo OSS →
NOVA
NOVA — Execution Layer
Workflow engine · Multi-tenant runtime · Cost tracking
▼ governed by
Matimo — Governance Layer
Policy registry · Tool RBAC · HITL approval · Audit trail
▼ runs on
⚙️
Enterprise Runtime
Cloud-hosted on AWS · Multi-tenant ECS Fargate · RDS PostgreSQL
Who uses NOVA

Built for the teams that
run the enterprise

NOVA is designed for three distinct roles — each gets what they need without stepping on the other.

AI / Automation Teams

Design agent workflows that actually ship

Use the visual studio to compose multi-step agent workflows, connect to enterprise systems, and deploy to production — without managing infrastructure or fighting governance requirements.

Engineering Teams

Extend via API without reinventing the stack

Full REST API access gives engineers control over agent configuration, workflow execution, and tool management — while inheriting multi-tenancy, audit trails, and governance by default. Skip the boilerplate. Build what actually matters.

IT / Security / Compliance

Control what runs, what it touches, and why

Matimo's policy engine classifies tool risk automatically — low-risk actions execute freely, high-risk actions route to a human approval queue. Every execution is logged with tenant context. Audit logs exportable for your SIEM.

Platform scale

Built for production
from day one

Key architectural capabilities designed around what enterprise deployments actually require.

16
Reasoning strategies — ReAct, CoT, ToT, Plan-Execute, Reflexion, Multi-Agent, and more. Select the right one per agent.
∞ tenants
Native multi-tenant isolation — no separate deployments required per customer or business unit
100%
Execution audit coverage — every tool call, every decision, every token tracked with tenant context
248+ tools
Tool actions across 8 Matimo providers — Slack, Gmail, GitHub, Notion, HubSpot, PostgreSQL, Mailchimp, Twilio. Plus any MCP server.
Enterprise Early Access

Ready to deploy your
first agent workforce?

NOVA is now open for early access. Sign up directly and start building your agent workforce — or schedule a walkthrough with the team if you'd like to see it in action first.

Sign up for NOVA → Schedule a walkthrough

Building on Matimo OSS already? Your policies carry over.