Sovereign AI for real operations

Intelligence
with a sense
of place.

AI that understands your work, remembers what matters, and helps your team act. One private agent engine for business workflows, digital services and technical operations.

Designed, installed & maintained by Nunki Labs.

THE EXECAI ENGINELOCAL FIRST
Intelligence with memory at its centreA conceptual constellation connects the ExecAI agent runtime to persistent memory, specialist agents, tools and model inference. MemoryExperience retained Specialist agentsA mission for each Tools & equipmentConnected to your site Your modelsLocal or chosen cloud
MEMORY. AGENCY. CONTROL.Conceptual view
YOUR WORK.
YOUR INTELLIGENCE.
Local first
Private inference, without API fees
Memory driven
Knowledge beyond a conversation
Human governed
Tools and permissions you define
A general-purpose agent engine

One engine.
Many fields of work.

From everyday office work to specialised industrial systems, ExecAI connects reasoning, knowledge and action around the way you operate.

Business & office

Make everyday work flow.

Build assistants around internal documents, prepare reports, organise information and automate repetitive steps across your existing business tools.

Document workflows · reporting · internal knowledge
Finance operations

Turn paperwork into prepared work.

Scope workflows for invoice data extraction, document classification, reconciliation preparation and accounting reports, with your team reviewing and validating the results.

Invoices · bookkeeping preparation · reconciliation support
Websites & customer care

Put intelligence where customers are.

Integrate knowledge-based chatbots, prepare support replies, route customer requests and assist website or catalogue updates. Connect stock monitoring to the systems that hold your inventory.

Customer support · website operations · inventory monitoring
Industry & remote assets

Understand the installation.

Connect equipment, energy systems and operational data to agents that inspect trends, explain anomalies and support supervision, maintenance and alerting.

Energy · machines & sensors · technical supervision

These are project directions to scope with you. The audit establishes the data, connectors, permissions and checks each workflow needs.

The local advantage

Useful intelligence.
Even beyond the network.

When a local model is capable enough for the job, it can handle the routine on your own hardware. A cloud provider becomes a project choice.

Continuity

Keep working offline.

A remote installation can keep measuring, reasoning and raising local alerts during an internet outage. Its models, tools and data stay on site.

Messages to a remote phone still need a working link.

Operating cost

No model API bill.

Local inference needs no model-provider API and incurs no per-request or per-token provider fees. A routine supervision task can run as often as needed.

Hardware, electricity and agreed maintenance remain part of the project cost.

Privacy

Keep sensitive data inside.

In a fully local workflow, prompts, documents, memories and measurements are processed on your infrastructure, without being sent to a model provider.

Your data policy determines which integrations and remote connections are allowed.

Local, cloud or both — decided after the audit. We assess the task, the model’s measured capability, connectivity, hardware, cost and privacy requirements. A hybrid project can keep routine or sensitive work local and use a cloud model for selected tasks where policy allows it.
Explore the deployment choices
The engine underneath

One coherent system.
Many capable agents.

ExecAI brings models, persistent memory, tools and specialised agents into a configurable runtime.

Meet the ExecAI engine
EXECAI / ENGINE MAPDownload diagram

Scroll horizontally to explore the complete map.

ExecAI engine architecturePeople and applications enter a configured instance. Its access rules surround an agent runtime and message bus. Agents use configured memory stores, allowed tools and model backends through an admission queue. A separate instance can communicate through a Redis bridge. Memory sources and model endpoints can be shared only when explicitly configured. People · console · applications · API INSTANCE · ACCESS & PROCESS IDENTITY Agent runtimeIdentity · profile · context · permissions Agent registry & message bus Specialist agentsMessages & delegation Another instanceOptional Redis bridge Persistent memoryHybrid search · local embeddingsPrivate stores · read-only corpusDated facts & session distillation Tools & proceduresEquipment · files · web · visionSkills · MCP tool serversExplicit allowlists per profile Models & admission queueLocal: Ollama · llama.cppCloud: configured providersCapacity · priorities · cancellation Conceptual architecture · sources, permissions and connections are configured per deployment.
Deployment boundary

Instances define the perimeter.

An instance is a running ExecAI service, with its own configuration, process identity, sessions and access rules. Deploy separate instances for separate trust boundaries; an agent profile alone is not tenant isolation.

The engine, demonstrated

Two field examples.
A much wider engine.

Argus and Hestia demonstrate the same general-purpose engine on real equipment. Their technical domains illustrate what the architecture can do; they do not define its limits.

PHYSICAL SECURITY

Argus.

A watchful eye.
A record you can question.

Turn camera events into descriptions, searchable history and alerts. Review the evidence and correct the agent’s verdict.

Explore Argus
Actual application · laboratory deployment
ENERGY & INFRASTRUCTURE

Hestia.

The whole installation.
A clearer picture.

Bring solar, batteries and consumption into one conversation. Follow trends, inspect cell health and anticipate reserve thresholds.

Explore Hestia
Actual application · laboratory deployment

Screenshots are recorded views of our systems, captured 17 November 2026. They are not a live feed.

Sovereignty, by design

Shape the work.
Choose the boundaries.

A document assistant, a customer-service workflow and an industrial supervisor can share the same engine. Each project selects its own tools, knowledge sources, model and access rules.

Understand the controls

Knowledge that stays useful

Search durable facts by meaning or exact terms. Keep private stores separate and share a read-only corpus deliberately.

Boundaries that are explicit

Configure who can enter, which tools an agent can use and what its process can reach.

A human with the controls

Inspect activity, set permission modes and stop work. Review outcomes against the evidence from your site.

From first question to daily use

Start focused.
Prove it in your workflow.

A practical engagement, with a result to review at every stage.

DISCOVER

Understand the work

Map the workflow, existing software and equipment, connectivity and privacy needs. Test whether a local model is capable enough, then agree on scope and data flows.

PILOT

Make it measurable

Test one use case with your real documents, systems or equipment. Compare results against an agreed measure before extending.

INSTALL

Make it yours

Integrate the system, document the installation and train the people who will operate it.

MAINTAIN

Keep it useful

Monitor, update and tune the system as the installation and its needs change.

Explore services & engagement
Field engineering, with a point of view

Built in the Carpathians.
Designed for your reality.

ExecAI is developed and operated by Nunki Labs in Pietreni, Romania, with R&B Risk Management. Our technical foundation comes from systems we operate ourselves. We adapt that architecture to your workflows, software and installations.

The people behind ExecAI
Let’s make it concrete

One workflow.
One good question.

A report that takes too long. Customers waiting for answers. A remote installation that needs to keep working. Tell us where useful intelligence would make a difference.

Talk about a pilotA conversation with the people
who build the system.