LoadingPage content is loading
The Agentic OS
CORTEX
A second brain for your operation. CORTEX is the agent runtime where reasoning, memory, tools, and policy come together — so autonomous agents do real work under rules you set.
Substrate chain
Pulse · Mesh · Synapse · Cortex · Aria · Forge
Six named systems on one runtime. Open any dock — the chain stays the same.
Skip immersive experience
Use the chapter track buttons, number keys 1 through 9, arrow keys, J and K keys, Page Up, or Page Down to move between chapters.
Loading scene
01 / 04 · Act I · Chatbox ceiling
Fracture
Most AI stops at the chatbox.
It summarizes, it suggests, it drafts — then a human still has to notice, decide, copy, paste, and do.
01Fracture · Act I · Chatbox ceilingWebGL
Swipe to explore
Continue into the platform.
Agent loop
How CORTEX reasons under gates
Sense
Pulse perception, Fabric records, and your own systems arrive as first-class inputs — agents reason over what is happening now, not over stale exports.
Pulse · Fabric · Systems
Reason
Mission templates bind tools, data scopes, and vertical behavior to a job — retention, franchise ops, brand protection — so judgment stays inside the lane.
Mission-bound
Gate
Autonomy levels decide whether the agent alerts, drafts, or acts. Severity rules and human checkpoints stand before irreversible work.
Policy · Human
Act
Approved work lands on the same operating record your console reads — every run leaves a reconstructable trail for the CORTEX console.
Audit trail
Stack context
Where the agent OS sits
CORTEX is not a chatbot bolted onto a dashboard — it runs on the same identity, records, and gates as the rest of the platform.
Layer 01 of 05
Identity
Triangle secure sign-on — one identity for every person and console on the network.
- Passkeys and secure sessions, built in from the start
- The public site and every operator console share one sign-on system
- The right brand context on every session — never a separate login per app
Gate
One secure sign-on is the onboarding gate — role-aware routing sends operators, partners, and counsel down the correct track.
Identity selected — Triangle secure sign-on — one identity for every person and console on the network.
Runtime
Agents that show their work
Powered by Pulse perception
CORTEX agents do not wait for a human to notice a problem. MCV Pulse streams typed events — sentiment shifts, narrative emergence, location divergence — straight into the agents responsible for acting on them.
Meet MCV PulseIdentity-scoped by MCV One
Every call an agent makes carries MCV One identity and consent context. An agent bound to three locations can never see the fourth — scopes are enforced at the platform layer, not by prompt etiquette.
How scoping worksSupervised in the CORTEX console
Every deployment ships with the CORTEX console — the operator surface where your team sets policy, reviews agent drafts, approves actions, and audits everything the system has done.
Request a walkthrough
The Agentic OS
Capabilities
Agents with real jobs
CORTEX agents are configured from templates with defined missions — retention, franchise operations, brand protection — each bound to the tools, data scopes, and vertical behavior its job requires.
Reasoning over live senses
CORTEX consumes MCV Pulse perception, Fabric connectivity, and your own systems as first-class inputs. Agents reason over what is happening now, not over stale exports.
Policy and human-in-the-loop gates
Every agent runs under explicit autonomy levels — alert-only, draft-only, or limited action. Severity-based rules decide when the agent acts, drafts, or waits for a human. Nothing external ships without the approval path you set.
Memory that compounds
Agents carry durable memory of decisions, evidence, and outcomes. Every escalation makes the next one smarter — institutional knowledge stops walking out the door.
How it works
Anatomy of a CORTEX agent
Template
Start from a pre-built role — membership retention, franchise operations, brand protection — or define your own mission.
Bind
Attach identity scope from MCV One: which locations, which data, which consent envelope.
Subscribe
Choose the events that wake the agent — a sentiment shift, a divergence alert, a narrative gaining velocity.
Govern
Set the autonomy level: alert-only, draft-only, or limited action — with severity thresholds and escalation paths.
Run
The agent perceives, reasons, drafts, escalates, and acts — every step logged, every claim tied to evidence.
Interface
Pulse → CORTEX Interface
Pull (Tools)
get_sentiment_summaryget_trajectorydetect_anomaliesget_evidencesuggest_response
+ Fit/Rest extensions
Push (Events)
pulse.sentiment.shiftpulse.narrative.emergedpulse.rest.divergence_detected
etc.
MCP Server for automatic tool discovery by any compatible agent runtime
Every call and event is scoped by MCV One identity + consent
Problem
AI that talks, in a business that acts
Most enterprise AI stops at the chatbox. It summarizes, it suggests, it drafts — and then a human still has to notice, decide, copy, paste, and do. The bottleneck was never intelligence. It was the missing operating system around it: identity, memory, tools, policy, and a way to act that a business can actually trust.
Meanwhile operational judgment does not scale. Your best district manager's instincts live in one head, cover eleven locations, and retire every time someone quits.
Human-in-the-loop
Trust is an architecture, not a promise
Autonomy without governance is a demo. CORTEX treats trust as plumbing: identity scopes enforced at the platform layer, human-in-the-loop gates keyed to severity, full audit trails of what each agent perceived, decided, and did. When an agent recommends a save play for a member at risk, the evidence package travels with it — quotes, velocity, trajectory — so the human approving it is actually informed.
Frequently asked questions
What is CORTEX?
CORTEX is the agentic operating system of the MCV platform — a runtime where AI agents with defined missions reason over live data, remember outcomes, and take bounded actions under human-set policy. It combines tools, memory, identity scoping, and audit in one layer.
How is CORTEX different from a chatbot or copilot?
Chatbots respond; CORTEX agents hold jobs. Each agent is configured with a mission, tool access, event subscriptions, identity scope, and autonomy limits. Agents act on live signals proactively — drafting responses, escalating with evidence — rather than waiting to be asked.
What is human-in-the-loop control in CORTEX?
Every CORTEX agent runs at an explicit autonomy level: alert-only, draft-only, or limited action. Severity-based gates decide when an agent may act, must draft and wait, or must escalate — and every decision and action is written to an audit trail.
Can CORTEX agents access all of my data?
No. Every agent call carries MCV One identity and consent context, enforced at the platform layer. An agent bound to specific locations or datasets cannot reach outside that scope — access limits are structural, not prompt instructions.
What tools can CORTEX agents use?
Agents use platform tools — Pulse perception queries, Fabric connectors into your existing systems — plus MCP-compatible tools from your own stack. Tool access is granted per agent template, so each role gets exactly what its job requires.
Does CORTEX replace my operations team?
No — it scales them. CORTEX agents handle continuous monitoring, first-draft responses, and evidence assembly, while humans keep approval authority on anything external. The goal is your best operator's judgment, present in every location at once.
Where does MCV Forge fit next to CORTEX?
Cortex thinks; Forge transforms. CORTEX reasons over live signals — predictive models, anomaly detection, generative recommendations — and MCV Forge is the adaptive transformation and outcomes engine that forges that intelligence into executable results: dynamic workflows and SOPs, personalized training paths, and playbooks that improve every cycle.
The substrate family
CORTEX is one of six named systems.
Pulse senses · Mesh connects · Synapse fires · Cortex thinks · Aria acts · Forge transforms
Cortex thinks — the central AI intelligence of the substrate: predictive models, anomaly detection, generative recommendations, optimization, and decision support, reasoning over everything the other systems carry.
Operator plane · product teaser
The substrate behind the marketing surface
Every department, approval gate, and service signal across the network lands in one operator shell — the same governed record the agents act on.
Control Room
VerticalsOpsIdentityGates
Network portfolio · live operations · service healthPROD
Verticals
5
Ops
Live
Identity
OK
Gates
Green
Live streamTeaser
Network graph · healthy
Triangle identity · session live
Payments loop · live
Deploy queue · idle
Next step
Schedule a briefing for your network
Corporate buyers, multi-unit operators, and investors start with a briefing — then a vertical path that fits the portfolio. Pilot one workflow, expand on evidence.