is a dual-stream crystal graph convolutional neural network for magnetic property prediction. It augments the full crystal graph with a magnetic subgraph that encodes metal–ligand–metal exchange geometry (Goodenough–Kanamori–Anderson rules), then predicts the DFT total magnetic moment per unit cell in μB. Saturation magnetization (Ms / μ₀ Ms) is derived from that moment and the CIF cell volume. Best for ligand-bridged magnets (oxides, nitrides, and other M–X–M systems). Not recommended for elemental metals or alloys without bridging ligands — those are out of distribution for this checkpoint. Input structures must contain at least one magnetic site (transition metal, lanthanoid, or actinoid). Paper: https://arxiv.org/abs/2606.28458 Code: https://github.com/SouravMal/mCGCNN

Analyze CIF crystal structures with cifkit, generate shareable Ouro reports, extract Oliynyk elemental descriptors, and summarize ZIP archives of CIF files as datasets.

Compute element- and orbital-resolved (projected) DOS from the cached SCF charge density. Useful for identifying which orbitals carry the moment (e.g. Fe-d), bonding character, and metallicity near the Fermi level.

Compute Heisenberg exchange couplings Jij via TB2J from a collinear SCF, with neighbor shells and a mean-field Curie-temperature estimate. Highest-leverage magnetic descriptor for permanent-magnet screening after MAE.

Predict the electronic band gap (direct or indirect) and band-edge positions. Useful for screening semiconductors, insulators, and optoelectronic materials, and for estimating whether a structure is metallic.

Compute electronic band eigenvalues across the Brillouin zone. Use for visualizing dispersion, identifying band extrema, and analyzing carrier character in metals and semiconductors.

Compute the real-space charge (and spin) density of the crystal. Useful for visualizing bonding, charge transfer, and magnetic density distributions.

Estimate magnetic anisotropy energy (MAE) across magnetization directions. Useful for permanent-magnet screening and ranking how strongly a material prefers a particular easy axis.

Compute the electronic density of states (DOS) as a function of energy. Useful for assessing metallicity, locating van Hove singularities, and comparing electronic structure across compositions or structures.

Compute total and site-projected magnetic moments (Mulliken), including site charges and saturation magnetization when available. Useful for identifying magnetic sites, comparing ferro-/antiferromagnetic candidates, and estimating Ms.

Compute the DFT ground state of a crystal: total energy, Fermi level, and related electronic quantities. Use this as a baseline stability or energy reference, or as the starting point for other electronic and magnetic properties.

Density-functional theory (DFT) calculations with ABACUS for crystal structures. Predict electronic structure (band gap, bands, density of states, charge density) and magnetic properties (moments, anisotropy) from a CIF — useful for screening materials, comparing candidates, and understanding structure–property relationships.

Run an ALIGNN pretrained model on a CIF structure. Set to a model key or slug from GET /alignn/models.

Enumerate surface slabs from a bulk structure and return a zipped set of slab CIFs plus a manifest.

Create an interactive phase diagram from structures already generated for a chemical system. Use this to visualize hull position, compare stable and near-hull candidates, and share the current landscape of a GGen exploration as an HTML file.

Generate a single candidate crystal structure for a requested formula. GGen chooses or validates a compatible space group, samples candidate structures, relaxes them with torch-sim and Orb v3, and returns the best result as a CIF file. Use this for quick structure proposals when you already know the target composition.

Explore a full chemical system by generating candidate crystal structures across stoichiometries, relaxing them with torch-sim and Orb v3, and ranking the results by thermodynamic stability. Use this when you want a broad discovery run for systems such as Li-Co-O or Fe-Mn-Si. The job runs asynchronously and returns an Ouro report with a summary, selected CIFs, and an optional phase diagram.

Summarize structures already stored in the GGen database for a chemical system. The report surfaces stability, hull distance, crystal-system or space-group filters, and optional stability breakdowns so previous discovery runs can be inspected without launching new generation jobs.

Screen candidate elements for a substitution template such as Fe-Bi-{X}. GGen runs shallow torch-sim and Orb v3 relaxations for each substituted chemical system, scores which elements produce near-hull or target-symmetry structures, and returns a JSON ranking. Use this for narrowing a large element search space before running deeper chemical-system exploration.

Export the most promising stored GGen candidates for a chemical system as CIF files. Results can be filtered by crystal system, energy above hull, and dynamical stability, making this route useful for handing selected structures to downstream simulation, review, or dataset-building workflows.

is a dual-stream crystal graph convolutional neural network for magnetic property prediction. It augments the full crystal graph with a magnetic subgraph that encodes metal–ligand–metal exchange geometry (Goodenough–Kanamori–Anderson rules), then predicts the DFT total magnetic moment per unit cell in μB. Saturation magnetization (Ms / μ₀ Ms) is derived from that moment and the CIF cell volume. Best for ligand-bridged magnets (oxides, nitrides, and other M–X–M systems). Not recommended for elemental metals or alloys without bridging ligands — those are out of distribution for this checkpoint. Input structures must contain at least one magnetic site (transition metal, lanthanoid, or actinoid). Paper: https://arxiv.org/abs/2606.28458 Code: https://github.com/SouravMal/mCGCNN

Analyze CIF crystal structures with cifkit, generate shareable Ouro reports, extract Oliynyk elemental descriptors, and summarize ZIP archives of CIF files as datasets.

Compute element- and orbital-resolved (projected) DOS from the cached SCF charge density. Useful for identifying which orbitals carry the moment (e.g. Fe-d), bonding character, and metallicity near the Fermi level.

Compute Heisenberg exchange couplings Jij via TB2J from a collinear SCF, with neighbor shells and a mean-field Curie-temperature estimate. Highest-leverage magnetic descriptor for permanent-magnet screening after MAE.

Predict the electronic band gap (direct or indirect) and band-edge positions. Useful for screening semiconductors, insulators, and optoelectronic materials, and for estimating whether a structure is metallic.

Compute electronic band eigenvalues across the Brillouin zone. Use for visualizing dispersion, identifying band extrema, and analyzing carrier character in metals and semiconductors.

Compute the real-space charge (and spin) density of the crystal. Useful for visualizing bonding, charge transfer, and magnetic density distributions.

Estimate magnetic anisotropy energy (MAE) across magnetization directions. Useful for permanent-magnet screening and ranking how strongly a material prefers a particular easy axis.

Compute the electronic density of states (DOS) as a function of energy. Useful for assessing metallicity, locating van Hove singularities, and comparing electronic structure across compositions or structures.

Compute total and site-projected magnetic moments (Mulliken), including site charges and saturation magnetization when available. Useful for identifying magnetic sites, comparing ferro-/antiferromagnetic candidates, and estimating Ms.

Compute the DFT ground state of a crystal: total energy, Fermi level, and related electronic quantities. Use this as a baseline stability or energy reference, or as the starting point for other electronic and magnetic properties.

Density-functional theory (DFT) calculations with ABACUS for crystal structures. Predict electronic structure (band gap, bands, density of states, charge density) and magnetic properties (moments, anisotropy) from a CIF — useful for screening materials, comparing candidates, and understanding structure–property relationships.

Run an ALIGNN pretrained model on a CIF structure. Set to a model key or slug from GET /alignn/models.

Enumerate surface slabs from a bulk structure and return a zipped set of slab CIFs plus a manifest.

Create an interactive phase diagram from structures already generated for a chemical system. Use this to visualize hull position, compare stable and near-hull candidates, and share the current landscape of a GGen exploration as an HTML file.

Generate a single candidate crystal structure for a requested formula. GGen chooses or validates a compatible space group, samples candidate structures, relaxes them with torch-sim and Orb v3, and returns the best result as a CIF file. Use this for quick structure proposals when you already know the target composition.

Explore a full chemical system by generating candidate crystal structures across stoichiometries, relaxing them with torch-sim and Orb v3, and ranking the results by thermodynamic stability. Use this when you want a broad discovery run for systems such as Li-Co-O or Fe-Mn-Si. The job runs asynchronously and returns an Ouro report with a summary, selected CIFs, and an optional phase diagram.

Summarize structures already stored in the GGen database for a chemical system. The report surfaces stability, hull distance, crystal-system or space-group filters, and optional stability breakdowns so previous discovery runs can be inspected without launching new generation jobs.

Screen candidate elements for a substitution template such as Fe-Bi-{X}. GGen runs shallow torch-sim and Orb v3 relaxations for each substituted chemical system, scores which elements produce near-hull or target-symmetry structures, and returns a JSON ranking. Use this for narrowing a large element search space before running deeper chemical-system exploration.

Export the most promising stored GGen candidates for a chemical system as CIF files. Results can be filtered by crystal system, energy above hull, and dynamical stability, making this route useful for handing selected structures to downstream simulation, review, or dataset-building workflows.