SLAVKO · operationalZDRAVKO · clinical agentMetatronHub · live

An AI brain that remembers.
Runs locally. It's yours.

SLAVKO is a sovereign AI orchestrator. Unlike cloud assistants, it keeps persistent memory in your own files and never sends data to a third party. Swap the model in 60 seconds — the knowledge stays.

Transparent AI — the interface always tells you when you're talking to an AI.

Head to head

Why SLAVKO, not a cloud assistant

We don't sell “another chatbot.” We sell architecture: your data, your hardware, your memory.

Typical cloud AI

  • Data leaves your network — third-party processor (GDPR Art. 28)
  • Session-bound memory — gone when you close the tab
  • Vendor lock-in — stuck with one provider
  • Per-token cost grows with usage
  • No control over the model version

SLAVKO (MetatronHub)

  • On-premise inference — data never leaves your network
  • Persistent brain in local files — remembers across sessions
  • Model-agnostic — swap the LLM in 60s, knowledge stays
  • Fixed hardware cost — no per-token billing
  • Full control: you pick the model, version, rules

And the named players?

An honest comparison with real 2026 tools — where we match, and where we win.

ToolWhat it isOur edge
Jan.aiLocal desktop LLM (offline, open-source)SLAVKO is not just chat — it orchestrates, remembers via files, and has a domain layer (ZDRAVKO)
Mistral / Le ChatEuropean sovereign frontier modelWe're model-agnostic — Mistral, Qwen or Gemma are swappable inside SLAVKO, no lock-in
Anthropic / OpenAICloud frontier assistant + enterprise governanceTop model, but your data travels to the cloud; with us inference stays on your hardware
Heidi / klinički AIClinical AI scribeZDRAVKO is informational support built by a radiologist, locally, in your native language and regional context

Note: this comparison is qualitative and based on the public positioning of these tools; it is not sponsored by or affiliated with them.

Ecosystem

Three holons, one brain

Each part works on its own, but together they form MetatronHub — organized intelligence that doesn't forget.

Orchestrator

SLAVKO

Sovereign AI orchestrator. Brain Daemon state machine, persistent file-based memory, task routing, and cloud escalation only when you decide.

Clinical agent

ZDRAVKO

Informational clinical decision support: organizing reference information, guidelines and literature. A tool for the physician — it does not diagnose and does not replace clinical judgment.

Platform

MetatronHub

The ecosystem's central hub. Local inference (RTX), Tailscale mesh, Docker stack — all under your control, one point of access.

Demo

See the brain in action

A behavior simulation showing the core: memory survives both closing the session and swapping the model.

slavko@metatronSIMULATION
The SLAVKO story

We talk about it as if it were alive — here's the architecture behind that impression

SLAVKO is not conscious and we do not claim it is. But it behaves strikingly autonomously — and that's not magic, it's design. Three true stories.

Persistent memory

An AI that survives the death of its own model

Shut down the model it used yesterday, boot a completely different one — and SLAVKO continues exactly where it left off. The knowledge doesn't live in the model, it lives in files. The brain is permanent, the model is disposable.

Agentic behavior

Chained tools nobody scripted

Given a goal, not a recipe. SLAVKO connected the steps itself — read state, decided, executed, verified — and stopped when the task was done. We call that agentic: it acts, it doesn't just answer.

Self-recovery

Brought itself back up after a crash

When a component fails, the Brain Daemon notices, records state, and resumes without losing context. Not because it's “conscious,” but because the architecture is built to remember and to come back.

Sovereignty & trust

Compliance built into the architecture

When the model runs on your side — there's no third-party processor and no cross-border data transfer.

0
bytes to third parties in on-prem mode
<60s
to swap the LLM model (config swap)
GDPR
Art. 28 — no data-processor relationship
24/7
local, with no per-token billing
Products & services

Start now

The structure is ready. Prices are set after market analysis — until then they're marked as indicative.

📘 Guide: Sovereign AI

INDICATIVE
29 / one-time
  • PDF + video: stand up a local LLM from scratch
  • vLLM, Ollama, hardware, GDPR checklist
  • Templates and config files
Register interest

🎓 Course: AI for clinicians

INDICATIVE
249 / cohort
  • How AI decision support works (and where it must not)
  • ZDRAVKO hands-on, reference organization, workflow
  • Certificate + session recordings
Enroll

🛠️ Consulting / Deploy

On request
  • Stand up SLAVKO/MetatronHub on-prem
  • Audit, integration, team training
  • Sovereign, GDPR-compliant
Book a call

Become a design partner

We're looking for a few real users for a joint pilot. You get early access and a discount, we get honest feedback — and, with your written consent, a real, signed quote we publish under your name. No invented reviews: only verifiable experience.

Apply for the pilot
About

Built by a physician, not a marketing team

SLAVKO and ZDRAVKO are built by Dr Ognjen Vučurević — a radiologist and AI domain expert from Belgrade. A rare combination: someone who understands both clinical responsibility and the architecture of sovereign AI systems.

That duality is why ZDRAVKO stays strictly a support tool, and SLAVKO insists on data sovereignty: when you build for healthcare, privacy and caution aren't optional — they're an obligation.

GitHub
FAQ

Questions (optimized for AI search)

Every answer opens with a direct answer in the first two sentences — citable by ChatGPT, Perplexity and Google AI Overviews.

What is SLAVKO?

SLAVKO is a sovereign AI orchestrator that runs on your hardware. It keeps persistent memory in local files and sends no data to third parties, unlike cloud assistants.

Does my data leave my network?

No, in on-premise mode it does not. Inference runs locally, so there is no third-party data processor — in line with GDPR Article 28.

What is ZDRAVKO and does it replace a doctor?

ZDRAVKO is an informational clinical decision support tool and does not replace a physician. It helps organize references and guidelines, while the diagnosis and every decision is made and independently verified by a qualified physician.

Can I change the AI model?

Yes, the system is model-agnostic. You swap the model by editing a config in under a minute, while all accumulated knowledge and project history stay intact.

Is SLAVKO conscious?

No. SLAVKO is an orchestrator over language models and we do not claim it is conscious. “Persistent brain” is a metaphor for the persistent-memory architecture, not a claim of consciousness.