GDBS (Geometric Database System) is a browser-native scientific computing SaaS platform. It delivers HPC-grade numerical precision across 8 physics domains using WebAssembly - no cluster, no installation, no job queue required. Every computation carries transparent drift and uncertainty tracking.

How does GDBS compare to ANSYS or COMSOL?

GDBS runs entirely in the browser via WebAssembly at a fraction of traditional simulation licensing costs. ANSYS Mechanical licenses start at $50,000+/yr with no browser access. COMSOL starts at $15,000+/yr per module. GDBS Standard starts at $25,000/yr ($2,083/mo) and Enterprise tops out at $75,000/yr ($6,250/mo), with 8 validated physics domains, transparent uncertainty tracking, no cluster requirement, and instant results - no job queue. HPC Lab is included with Enterprise. Numerical-relativity (BSSN + Z4c), gravitational-wave (LIGO + NRSur), and GR-MHD (GRMHD) addons are licensed separately at $60,000/yr ($5,000/mo) each on Standard or Pro, and are included free with Enterprise.

Does GDBS require installation or a cluster?

No. GDBS runs entirely in the browser using WebAssembly (Wasm). No installation, no cluster, no GPU, no job queue. It works on any device - phone, laptop, or workstation - with identical precision results.

What physics domains does GDBS support?

GDBS includes 8 physics domains: Quantum (energy levels, H2 ground state), Fluids (CFD, Taylor-Green vortex), Plasma (Landau damping, MHD, tokamak stability), Materials (bulk modulus, crystal symmetry), Geophysics (gravitational field, seismology), Ballistics (hypersonic aerodynamics), Cosmology (galaxy rotation, CMB), and Molecular (drug screening, protein stability).

How accurate is GDBS compared to supercomputer results?

GDBS reproduces published scientific references where one exists: LIGO's GW150914 final black-hole mass to 0.13% and final spin to 0.65% (Abbott et al.), GW170817 chirp mass to 0.3%; quantum-chemistry energies within 0.5 to 2% of the Szabo and Ostlund reference; the Blasius boundary-layer value 0.4696 within 0.002; numerical-relativity self-convergence of order 3.946; a magnetized accretion torus on GPU holding magnetic-field divergence to 2.37e-4; and the analytic Hawking temperature to 0.27% across 60 orders of magnitude. Results are reproducible live in the browser.

Is there a free version of GDBS?

Yes. GDBS has a permanent free tier covering the GDBS database and Theoretical Foundations module (citation required). A 3-day full trial unlocks all physics domain modules with no credit card required. Academic plans start at $99/yr (per 5 users, .edu required). Commercial tiers are Standard ($25k/yr or $2,083/mo), Pro ($50k/yr or $4,167/mo), and Enterprise ($75k/yr or $6,250/mo). HPC Lab and the BSSN, LIGO and GRMHD addons are included free with Enterprise; on Standard or Pro each addon is $60k/yr ($5k/mo), or $120k/yr ($10k/mo) for all three. Pricing subject to change at any time.

Now live - GDBS v1.0.0 · Windows x64

The database that
knows its own uncertainty.

GDBS is a geometric database system with HPC-grade precision computing built in. Every computation carries its error. Every result is verifiable. No black boxes. No hidden drift.

Open GDBS Lab Download Free - Win x64

Hawking radiation error 0.27%

GeoNum drift 0.345 shades

60 orders of magnitude validated

IEEE 754 outperformed

// GeoNum precision computation
const result = GeoNum.compute(hawking);
// → T_H = ℏκc / (2πk_B)
value: 6.169 × 10⁻⁸ K
drift: 0.345 shades
error: 0.27%
range: 10⁻³⁴ → 10⁻¹⁴

                        // IEEE 754 fails at 10⁻³⁴ scale

                        // GeoNum holds. Drift tracked. Verified.
                    

The Core Innovation

What the engine computes, and how it is checked.

Standard databases store data and return it. GDBS stores data geometrically - every value knows where it lives in precision space, how much it drifted to get there, and what its uncertainty is. That's not a feature. That's a different physics.

0.345

Shade drift on the Hawking-temperature chain spanning 10⁻³⁴ to 10⁻¹⁴, well under the 1.0 trust threshold; the result holds 0.27% accuracy vs the analytical closed form.

0.13%

Gravitational waves - reproduces LIGO's published GW150914 final black-hole mass, checked against Abbott et al. (LIGO/Virgo).

0.4696

Blasius boundary layer - the canonical flat-plate wall-shear value to within 0.002, checked against Falkner & Skan (1931).

Run the Theory Lab

GeoNum Precision

2048-shade zone/shade architecture. Drift compartments track error at every multiplication step - transparently, not silently.

Domain-Tuned Zones

One GeoNum class, domain-specific zone configurations. Theory, Quantum, Fluids, Plasma, Materials, Geophysics, Ballistics.

Seed-Based Storage

8-byte content addressing. 5D geometric positioning in semantic space. Automatic relationship discovery between data nodes.

Transparent Uncertainty

getUncertainty() and getRelativeUncertainty() on every result. No black box. Every computation auditable, every drift visible.

One geometric computing stack

One system. Three layers.

A unified geometric computing stack: precision, search, and reasoning. Precision (GeoNum) - Search (SyncSearch) - Reasoning (CORA), all on GDBS.

C.O.R.A.

Coordinated Ontological Referencing Architecture

The intelligence node. A browser-native AI assistant with persistent geometric memory - knowledge is stored in GDBS, not in a context window. Responses are grounded, traceable, and domain-aware.

Persistent GDBS memory - knowledge survives sessions

Domain-aware reasoning - physics, materials, geophysics

Native gdbsio engine - 4.7M-seed corpus, sub-10ms retrieval

Voice synthesis - natural TTS, hot-reload modules

Open C.O.R.A.

SyncSearch

High-Speed Geometric Discovery Engine

Enterprise search without the cluster. GDBS geometric positioning replaces the inverted index - semantic proximity in 5D space means relevance is a geometry problem, not a keyword problem.

Sub-millisecond queries - no Java, no cluster overhead

90% cost reduction - vs Elasticsearch at comparable scale

Semantic + full-text - GDBS proximity + FTS5

Embeddable - single binary, add to any stack

Request Early Access

Protected by U.S. Patent App. No. 63/909,979

GeoNum Engine

Drift-tracked precision, command line

The precision core of GDBS as a licensed CLI. Every result carries a drift figure and a trust verdict derived from the engine's own propagation - so you know whether to rely on a number, not just what it is.

Drift-tracked - precision carried through the calculation, not assumed

Trust verdict - PRECISE / DRIFTING / UNRELIABLE / ERROR on every result

Single native binary - runs locally, no cluster, no service

3-day trial - then $60 / $4k / $8k per month, cancel anytime

Get GeoNum Engine Open

Validated Precision

Measured against published references.

Each result below is computed by the engine and compared against a published scientific reference: a detection paper, a textbook value, or a convergence theorem. Platform values are reported as measured; references are cited as published. Nothing here is a curve fit.

Run Live Benchmarks

0.13%

Gravitational-Wave Merger

Reproduces LIGO's published GW150914 final black-hole mass to 0.13% (spin to 0.65%, chirp mass to 0.4%), checked against Abbott et al. (LIGO/Virgo). Computed and compared, not fitted to the detection.

3.946

Numerical Relativity

The spacetime (BSSN) evolver is fourth-order accurate. The order falls out of grid refinement - measured 3.946 against the theoretical 4 on the standard Apples-with-Apples testbed, not asserted.

0.27%

Hawking Temperature Precision

T_H reproduced across 60 orders of magnitude (10⁻³⁴ to 10⁻¹⁴) to 0.27% of the analytical closed form, GeoNum drift 0.345 shades. IEEE 754 degrades silently at this scale; GDBS tracks every digit.

$75k

vs $M/yr Cluster Time

HPC-grade precision at a fraction of cluster cost. Enterprise tops out at $75k/yr - full platform, HPC Lab, dedicated SLA. No job queue. Instant results. Transparent uncertainty on every output.

Benchmark Results

GDBS vs Published HPC Reference

Every result is reproducible in your browser. No cluster access required.

Benchmark	GDBS Result	Published Reference	Error	Source
GW150914 final black-hole mass	63.02 M☉	63.1 M☉	0.13%	Abbott et al. (LIGO/Virgo), PRL 116
GW150914 final spin a_f	0.6855	0.69	0.65%	Abbott et al. (LIGO/Virgo)
GW170817 chirp mass	1.185 M☉	1.188 M☉	0.3%	Abbott et al. (LIGO/Virgo), PRL 119
Quantum chemistry (H₂, H₂O, CH₄)	within tol.	STO-3G HF	0.5-2%	Szabo & Ostlund (textbook HF)
Blasius flat-plate wall shear f″(0)	0.4696	0.4696	<0.002	Falkner & Skan 1931 / Schlichting
Numerical relativity self-convergence	order 3.946	4 (theorem)	4th-order	Apples-with-Apples gauge testbed
Hawking temperature T_H precision	0.27%	analytical closed form	0.27%	60 orders of magnitude (10⁻³⁴-10⁻¹⁴), drift 0.345 shade
GPU constrained transport, \|div B\|	2.37e-4	< 2e-3 gate	held	Magnetized accretion torus, evolved on GPU

Reproducible against the engine's own release-gating test suites - gravitational-wave detection, numerical-relativity convergence, density-functional chemistry, fluid boundary layers, and precision chains - at gdbs.getvaultsync.com.

Access Tiers

Start Free. Scale to Research.

From exploratory tools to full HPC-grade research workflows. No credit card required to start.

Free / Trial

Free

3-day full trial on signup. No credit card.

GDBS database

Theoretical Foundations

All modules (3 days)

Create Account

Academic

$99-$3.5k/yr

.edu required. Per 5 users. Student $99 · Faculty $249 · Lab $3,500.

All physics domains

HPC Lab + DFT Engine

Multi-physics coupling

Grant-friendly pricing

Standard / Pro / Enterprise

$2,083-$6,250/mo

Commercial physics modules, multi-physics coupling, SLA. HPC Lab + all addons on Enterprise. BSSN, LIGO, GRMHD sold as separate addons. ($25k-$75k/yr annual)

Standard / Pro physics domains

Multi-physics coupling + batch sweeps

Priority / dedicated support

Enterprise: HPC Lab + all addons + source code option

How it compares

GDBS vs Traditional Simulation Licensing

Tool	License / yr	Precision tracking	Browser-accessible	Cluster required
ANSYS Mechanical	$50,000+	None	No	Often
COMSOL Multiphysics	$15,000+	None	No	Optional
VASP (ab initio)	$5,000+ academic	None	No	Required
HPC Cluster Time	$1M+/yr	Post-hoc only	No	Is the cluster
GDBS Standard	$25,000/yr	Transparent drift	Yes	No
GDBS Enterprise	$75,000/yr	Transparent drift	Yes	No

Research Program

We're accepting a limited cohort of research institutions and engineering teams for full research access. Participants get the complete platform - Multi-Species MD, coupled MD↔FEM, batch sweeps, and HPC Lab - at no cost during the program.

Requirements:
· Institutional or academic affiliation
· Cite GDBS in any publications using results from the platform
· Share anonymised compute logs to improve precision validation
· Brief monthly check-in with the GDBS team

Academic and student pricing available - contact us directly. We respond to everything.

Apply for Access

The database that
knows its own uncertainty.

What the engine computes, and how it is checked.

GeoNum Precision

Domain-Tuned Zones

Seed-Based Storage

Transparent Uncertainty

One system. Three layers.

C.O.R.A.

SyncSearch

GeoNum Engine

Measured against published references.

Gravitational-Wave Merger

Numerical Relativity

Hawking Temperature Precision

vs $M/yr Cluster Time

GDBS vs Published HPC Reference

Start Free. Scale to Research.

GDBS vs Traditional Simulation Licensing

Research Program

See the precision for yourself.

Talk to us.

COO

CLO

Company

The database that knows its own uncertainty.

What the engine computes, and how it is checked.

GeoNum Precision

Domain-Tuned Zones

Seed-Based Storage

Transparent Uncertainty

One system. Three layers.

C.O.R.A.

SyncSearch

GeoNum Engine

Measured against published references.

Gravitational-Wave Merger

Numerical Relativity

Hawking Temperature Precision

vs $M/yr Cluster Time

GDBS vs Published HPC Reference

Start Free. Scale to Research.

GDBS vs Traditional Simulation Licensing

Research Program

See the precision for yourself.

Talk to us.

COO

CLO

Company

The database that
knows its own uncertainty.