X‑HydrideLab
X‑HydrideLab

The Grok-native discovery platform for hydride superconductors.

X-Hydride Lab extends the Grokene framework into structured candidate generation, weighted X-Score evaluation, simulation preparation, and audit-grade provenance. Every output is an exploratory hypothesis intended to enter a rigorous computational and experimental validation pipeline.

Open LabView Framework

Lineage

From Grokene to X-Hydride.

X-Hydride Lab inherits Grokene's discipline and extends it to the rigor required by hydride superconductor research.

00

Grokene

First proof-of-concept for AI-guided superconducting candidate discovery. Established the structured prompt → candidate JSON → audit hash pipeline.

01

X-Hydride Lab

Extends the Grokene workflow into hydride-based superconductors with stricter scoring, simulation handoff, and a provenance ledger.

Framework

A discovery pipeline you can audit.

Inputs feed candidate generation; validation gates govern progression. Every artifact in between is hashed and recorded.

Figure 01Discovery pipeline · inputs (top) feed candidate generation; validation gates (bottom) gate progression.
I01Composition family
I02Target pressure
I03Tc range
I04Stability preference
  1. 01Input Hypothesis

    Research objective

  2. 02Grok Candidate Generation

    Strict JSON schema

  3. 03X-Score Evaluation

    8 weighted axes

  4. 04Simulation Template

    QE / DFPT / EPW

  5. 05Research Note

    Cautious draft

  6. 06Provenance Hash

    SHA-256 ledger

G01DFT relaxation
G02DFPT phonon
G03EPW Eliashberg
G04Experimental

Workflow

Five disciplined steps.

Generate, score, prepare, document, anchor. Each step produces a reviewable artifact.

  1. 01

    Generate

    Submit a structured research request. xAI Grok returns strict-JSON candidate hypotheses constrained by an explicit schema.

  2. 02

    Score

    Eight weighted subscores produce a single overall X-Score. Risk flags surface phonon, thermodynamic, and pressure concerns.

  3. 03

    Prepare

    Generate CIF, Quantum ESPRESSO, GPAW/ASE, DFPT phonon, EPW, and convergence templates as a deterministic bundle for review.

  4. 04

    Document

    Draft a cautious academic-style research note framed as a hypothesis requiring DFT, DFPT, EPW, Eliashberg, and RPA validation.

  5. 05

    Anchor

    SHA-256 hashes for input, output, report, and simulation artifacts are written to a provenance ledger ready for on-chain anchoring.

Validation

Hypothesis, not claim.