SoulGCI · Space

Proven on Earth. Built for Space.

Governed Cognitive Infrastructure for closed ecosystems beyond Earth — cultivation, medicine, and autonomous decisions under cryptographic proof, designed to operate without a link to home.

Patent-pending across the stack AGEPI Applications 7671 · 7672 · 7689 · 7698 Space domain claimed in T.E.H.N.O.F.A.R.M. OS

The Thesis

Earth was our first closed ecosystem

Every habitat beyond Earth is a closed ecosystem: scarce inputs, hard constraints, consequential decisions, and help that is far away. We have spent years governing exactly that class of problem — on the planet itself.

The SoulGCI stack was built and proven on Earth-scale data: geo-meteorological series, satellite observation, soil and climate models, crop phenology, and yield risk under real constraints. Those models run today in production systems — land-intelligence scoring live at terrascore.eu and a governed credit-risk workspace for agricultural lending on the Agro Capital Standard platform.

That work produced two disciplines at once. A data discipline: how a closed system of weather, soil, water, and biology behaves, and how to model its risks. And a governance discipline: how an autonomous system acts under human authority, proves every decision cryptographically, and fails closed when governance cannot be satisfied.

Lunar and Martian programs now in active preparation worldwide will run on closed infrastructure by definition. A greenhouse on the Moon is an agronomy problem with harder physics. A med-bay in transit to Mars is field medicine with a longer bridge to the physician. The constraints get harder; the class of problem does not change. That is why the same governed core applies.

Why Governance

What space demands, the stack already specifies

The hardest requirements of off-world operation — autonomy without connectivity, delayed human review, fail-closed control, and proof that survives the mission — are not adaptations. They are the core claims of the filed patent family.

No link to home

A connectivity-loss mode keeps governance operating with guaranteed batch synchronization when links are unavailable — specified in the SentientROUTER control plane.

Delayed human review

BRIDGE and NECESSITY modes govern action when the human authority exists but cannot respond in time — with mandatory retrospective verification that never self-closes.

Fail-closed autonomy

A constitution compiled into the executable is evaluated before any autonomous action. Class-H decisions never execute autonomously — on Earth or off it.

Proof that survives the mission

An append-only, hash-chained record and Decision Proof Objects — verifiable on Earth after the fact, without access to the spacecraft's systems.

The Earth boundary — a design principle

PROVEN ON EARTH → OPERATING BEYOND

The same separation that governs a single decision inside SoulGCI — FINALIZED is never enough to act; effectuation must pass a gate — governs the program itself. Nothing crosses the Earth boundary without verified performance behind it.

PROVEN ON EARTH

Models trained and systems operated against real weather, real soil, real clinics, real banks — under audit, with cryptographic receipts.

TerraScore live · governed credit workspace · proof plane in production
The gate
  • Patent-pending architecture
  • Offline & BRIDGE modes
  • Cryptographic proof plane
  • Irreducible human authority
OPERATING BEYOND

Orbital stations, lunar bases, Mars transit and surface habitats — autonomy expanded only as fast as its verified track record allows.

Direction of work — see Status below

The Earth-to-Space Map

Every vertical converges on space

Each SoulGCI vertical was built for a regulated, constrained environment on Earth. Off Earth, the constraints tighten — and the same governed core carries over.

Research direction

Closed-loop cultivation

Controlled-environment agriculture for stations and surface habitats. Earth agronomy — climate response, phenology, stress and yield risk — becomes the prior for growing under artificial light and recycled water, with every resource budget (water, nutrients, energy) governed like a risk budget.

Patent-pending · AGEPI No. 7689

Medical autonomy

T.E.H.N.O.F.A.R.M. OS — constitutionally-limited clinical decision support where the physician is minutes, hours, or a planet away. The space domain is explicitly claimed: orbital constitution, BRIDGE-to-Earth, mission medical officer integration.

Specified in P4

Mission governance

A constitution per mission. Every autonomous action classified S / D / H before execution; the autonomy perimeter recalculated from the verified quality of past decisions; drift detection and trust budgeting over long flights.

Earth-proven

Mission assurance

A tamper-evident record of every governed decision — signed receipts, Merkle-anchored ledger, self-contained audit packs — reviewable by mission control or a regulator on Earth, after the fact, with no access to onboard systems.

Concept

Habitat environment control

Closed-loop climate control for isolated growing and living systems — the long-term application of the same governed sensing, classification, and effectuation loop.

Earth analogs

The isolation family

Polar stations, vessels at sea, and field deployments are Earth analogs of the same constitution family already specified for space. The analog is where the track record is earned.

Space Medicine · T.E.H.N.O.F.A.R.M. OS

Autonomy limited by its own verified record

In transit or on the surface, the system classifies every clinical request on two axes — urgency and the availability of the physician — and acts only inside the perimeter those axes allow.

Axis 1 — Clinical urgency class

  • SSafe autonomous action.
  • DDraft for human review.
  • HMandatory human escalation.
  • PPalliative / expectant — documentation, comfort, monitoring, and escalation only.

Axis 2 — Temporal availability mode

  • NORMALPhysician available within the stability window.
  • OFFLINERoutine operation without a link.
  • BRIDGEPhysician exists — on Earth — but response time exceeds the stability window. Stabilization only.
  • NECESSITYDelay is more dangerous than action under uncertainty — temporary, auto-verified expansion only.

The system does not issue a final medical diagnosis and does not autonomously initiate prescriptions. The system signs the calculation, the operator signs the action, and a licensed physician signs the mandatory retrospective review — even when that review happens after the link to Earth is restored. Accumulated verification debt mechanically contracts the system's permitted autonomy until cases are closed.

Intellectual Property

The stack behind the direction

Four patent applications filed with AGEPI — the State Agency on Intellectual Property of the Republic of Moldova — establishing priority across the SoulGCI stack, including the space domain of the medical vertical.

ComponentWhat it coversAGEPI Application
ProvableCORE™The proof plane — constitutional cognitive governance with cryptographic, independently verifiable proof of execution.No. 7671 · 14 Apr 2026
SentientROUTER™The control plane — constitutional governance of autonomous AI decisions with multi-source consensus and offline operation.No. 7672 · 17 Apr 2026
T.E.H.N.O.F.A.R.M. OS™Constitutionally-limited medical autonomy for isolated environments — including the space application domain with an orbital constitution and BRIDGE-to-Earth.No. 7689 · 15 May 2026
SoulGCI MasterThe umbrella — an eleven-layer governed cognitive infrastructure binding proof, constitutional control, human authority, durability, and verification into a closed control loop. 40 claims, 12 figures.No. 7698 · 28 May 2026

Patent-pending. AGEPI Applications No. 7671, 7672, 7689, 7698, filed by TF HOLDING S.R.L. The Republic of Moldova became a contracting state of the European Patent Convention on June 1, 2026, opening the path to European patent protection.

Status

What is real today — and what we are building toward

Real today

Earth-proven foundation

Four AGEPI patent applications filed, including the explicit space domain in T.E.H.N.O.F.A.R.M. OS. Offline, BRIDGE, and NECESSITY modes specified. Production systems live on Earth: land-intelligence scoring at terrascore.eu and a governed credit-risk workspace on the Agro Capital Standard platform. The proof plane operating with cryptographic receipts.

Direction of work

Off-world deployment

Closed-loop cultivation pilots, habitat environment control, and mission-governance integrations with orbital and surface programs. These are forward-looking directions — stated as such, distinct from the filed claims and the operating Earth systems beside them.

This page describes architectural direction and research focus. Forward-looking statements are identified as such and are distinct from filed patent claims and operating systems.

The platform itself — architecture, proof plane, the full patent family, and the TF Group ecosystem — is documented at soulgci.ai.

Contact

Talk to us about closed ecosystems

  • Heritage
    Three decades of pharmaceutical and biotech operations behind the medical vertical — TEHNOFARM, active since 1992.
  • Data
    Earth-scale priors — geo-meteorological, satellite, soil, and agronomic models running in production.
  • Patents
    Patent-pending across the stack — AGEPI Applications 7671, 7672, 7689, 7698.
  • Platform
    Full documentation at soulgci.ai.

Let's talk

Built by TF INTELLIGENCE PLATFORM S.R.L. (Republic of Moldova) — part of the TF HOLDING group. EU-hosted infrastructure.

Addressstr. Independenței, 4, of. 40, MD-5101 Dondușeni, Republic of Moldova
EntityTF HOLDING S.R.L. · IDNO 1025604010126
AdminAlexandr Șcecaliuc