Operational research prototype — live since 28 July 2026
// Live-State Authentication

ZEUS X-Trust

We stopped storing conventional secrets. The password does not unlock a vault — it drives a living neural network into a state that either exists, or does not.

ZEUS X-Trust — The only solution for the treasure that counts.™

For investors → See the measured evidence
6–11
orders of magnitude separation
0.0
correct-state drift — bit-exact
18
layer live neural network
0
stored static keys
ACCEPT
DENY
// The mechanism

Authentication without a stored key

ZEUS X-Trust does not depend on a conventional static authentication key stored somewhere for later comparison. The valid authentication relation is generated live — and only reproduces inside the enrolled instance.

01

Password

The password is the only input the user provides. Nothing derived from it is stored for later comparison.

02

Positional map

A password-controlled positional map addresses specific neurons in specific layers of the network.

03

300 iterations

The live neural network runs 300 iterations per authentication and converges into a stable amplitude state.

04

Amplitude profile

The instance-specific amplitude state created during enrollment is compared against the live state.

05

Verdict

Accept only when map and amplitude profile both match. Otherwise: deny.

Two independent separations at once: a wrong password fails at the positional map — a newly initialized instance fails at the amplitude profile. Only the correct password-controlled map inside the enrolled live instance satisfies both conditions.
// Measured results

From experimental validation to a running prototype

Every figure below comes from the documented validation runs across Stages 3 to 5 and the first end-to-end prototype test.

Stage 3 — Plateau stability & distinguishable maps

4 architectures · 400 complete runs · 4,000,000 network iterations

Four original FRN architectures — Original, Pyramid, 4weights, Extended — each with 10 passwords, 10 genuine re-initializations per password and 10,000 iterations per run. Original scripts unmodified, full positional amplitude maps stored and evaluated.

✓100/100 runs passed Test A for every architecture
✓10/10 password groups passed Test B for every architecture
✓Mean intra-window drift: 0 — control window max_diff 0.00000000
✓No NaN, no Inf, stable late-state plateau throughout
✓Genuine re-initializations produced distinguishable positional amplitude maps
✓Euclidean distance as primary metric, Pearson correlation as secondary reference
Key insight: large transient amplitudes are not instability if the final positional amplitudes converge to an exact and persistent plateau. The decisive object is the complete positional amplitude map — which neuron, in which layer, holds which exact amplitude after convergence.
Stage 4 — Profile vectors & map binding

Binding deterministic password maps to stable live-state amplitude profiles

Validated across all four original FRN architectures, including a separate shared-initialization experiment that isolated the password effect from random initialization.

✓100/100 valid same-instance references accepted
✕900/900 wrong-password maps rejected
✕900/900 obsolete references rejected after re-initialization
✓Deterministic password maps across all tested architectures
✓Bit-exact stability within the established live-state plateau
✓0 of 450 password pairs per architecture were identical
Identical architecture, identical initial weights, only the password changed — and different passwords produced different stable full-network states. This supports the dual role of the password: it defines the map that addresses neural positions, and it influences the stable live state generated from an otherwise identical initialization.
Stage 5 — Integrated authentication logic

Full-run results across 80 individual runs

The complete live-state authentication workflow: correct-login acceptance, rejection of incorrect passwords, instance-specific amplitude-reference binding and protection against reference overwriting during authentication.

✓40/40 correct logins accepted
✕360/360 incorrect-password attempts denied
✓0/40 enrollment references modified
✕40/40 newly initialized instances denied against the old reference
✓0 NaN or Inf failures across 80 individual runs
✓Combined prototype config (Pyramid + 4weights): 10/10 accepted, 90/90 denied
Scope note by the author: Appendix A reports the controlled experimental validation of Stages 3–5. It does not claim independent certification, production deployment, sustained operational exposure, or completion of the Stage 6 adversarial programme. External third-party replication is a separate verification layer.
Prototype — first end-to-end test

A working local application with a real 18-layer neural network

Not a simulated interface element. File sealed, authentication gate live, attack variants tested against the enrolled instance.

✓File successfully sealed · correct password: ACCEPT
✓32/32 mapped positions matched
✓Correct-state drift 0.0 — bit-exact
✕Anagram attack: DENY
✕Single-character substitution · truncation · case mutation: DENY
✕Unrelated password: DENY · no NaN or Inf failures
The anagram attack is the revealing case: same characters, different order — still rejected through the live-state component rather than through a simple character-set comparison. Measured separation between the valid state and every tested invalid state: 6–11 orders of magnitude of amplitude separation, with the four-decimal verification window lying inside the empty numerical space between those conditions.
6–11
Orders of magnitude
Amplitude separation between the valid state and every tested invalid state
4,000,000
Network iterations
Across 400 complete Stage 3 validation runs on 4 architectures
1,300+
Denied attempts
900/900 wrong-password maps, 360/360 wrong passwords, 40/40 stale instances
0
NaN / Inf failures
Numerically clean across every documented validation run

Restart kills the vault — volatility as tamper evidence

The prototype deliberately uses a volatile neural instance. A restarted instance produces a new amplitude realization, the previous enrollment reference becomes invalid and a new enrollment is required. What looks like a limitation is the security property: there is no persistent secret left behind to steal, and any tampering with the instance destroys the authentication relation it was meant to unlock.

No stored static key

There is no key object at rest that can be exfiltrated, cracked offline or recovered from a backup. The relation is produced live or not at all.

nothing at rest

Post-quantum by construction

The attack surface is not a mathematical one-way function but a live amplitude state inside an instance-specific network — there is no algebraic shortcut to invert.

no invertible target

Watchdog validation

An independent watchdog network is in implementation, moving the system from request-based authentication to continuous live-state integrity monitoring.

in implementation
// Development path

Where the prototype stands — and what comes next

This is a research prototype. The protected file currently remains unencrypted behind the authentication gate: the present experiment validates the live-state authentication mechanism itself, not yet the complete storage-security layer. That distinction is deliberate.

~1,600 h
Human research time invested
~200 GB
Validation data generated
~75 M
Tokens for X-Trust in 4 weeks
~7 h
From design to running prototype
Stages 3–5 validated
Plateau stability, map binding and integrated authentication logic documented in Appendix A.
Operational prototype live
Local application with a real 18-layer network, first end-to-end test passed.
Removal of research-only password storage
Eliminating the last research-grade convenience from the implementation.
Controlled attack campaign — ~500 scenarios
Testing whether the surrounding implementation can be bypassed, manipulated or compromised.
Independent watchdog implementation
Continuous live-state integrity monitoring instead of request-based authentication.
Encrypted storage integration
Completing the storage-security layer behind the authentication gate.
Large adversarial campaign — 20,000+ executions
More than 20,000 distinct attack executions and variants, with published evaluation material.
// The system behind the system

Built by an agent architecture, directed by one scientist

The operational prototype was not produced through a conventional software-development process. It was designed, implemented, tested and consolidated inside a local agent architecture — with research direction, architecture and scientific interpretation remaining with Stefan Trauth.

A
Apollon
routes
M
Medusa
plans
P
Pandora
builds
K
Kosmos
plans the outcome

Apollon routes. Medusa plans. Pandora builds. Kosmos plans the outcome. All four are a product of GeoFold AI — a deep-tech venture by Trauth Research LLC.

// Investment

Open for venture capital and strategic investors

ZEUS X-Trust is a prototype that could introduce a new class of cryptography. From this point on, GeoFold AI is actively seeking venture capital and other investors to move it from validated prototype to hardened product.

What is on the table

  • ▸A working, measured prototype — not a concept paper. Every figure on this page comes from documented validation runs.
  • ▸A published preprint and an experimental validation appendix, openly citable on Zenodo.
  • ▸A defined adversarial programme: ~500 controlled scenarios, followed by 20,000+ attack executions and variants.
  • ▸Target markets: agentic operating systems, AI-controlled infrastructure, research institutions, specialised security providers and infrastructure operators.
  • ▸Proprietary source code and implementation retained; methods and selected raw data released under controlled access.
  • ▸Technical due diligence and live demonstration available on request, under NDA.
Request the investor briefing →

Direct email only — no contact form, no tracking, no data collection on this site.

Terms & Conditions

All services are B2B only. No products or services are sold via this website; it serves informational purposes. Trauth Research LLC retains full intellectual property rights for all developed models, structures, and technical methods unless otherwise agreed by written contract.

Governing law: Federal Republic of Germany. Place of jurisdiction: Berlin.

Liquidated Damages
Unauthorized replication, reverse engineering, or use in GPU clusters, data centers, or military contexts: USD 5,000,000,000 (corporate) · USD 10,000,000,000 (governmental). Accessing this system constitutes acceptance of these terms.
Appendix A — Experimental Validation
From stable amplitude states to an integrated live-state authentication prototype · Zenodo
↗
Preprint — ZEUS-Derived IGAN
AI-powered quantum-resistant authentication, key management and live-state integrity · Zenodo
↗
Research log
Stage-by-stage publication of methods and results · LinkedIn
↗
The research behind the venture
Trauth Research LLC
Independent research on informational ontology, emergent geometry and post-quantum AI. GeoFold AI is the deep-tech venture that turns that research into products.
Visit trauth-research.com →