Preparing an atomistic potential evaluation
Plan a source-grounded evaluation of MACE, NequIP, SchNetPack and TorchANI by defining reference chemistry, checking model coverage, aligning inputs and outputs, and documenting reproducibility conditions.
Collections
Themed collections of chemistry AI resources.
Chemical data interfaces for agent workflows
Compare PubChem and ChEMBL community MCP integrations, the Materials Project API client and Scientific Agent Skills, then plan provenance, permissions and failure checks before connecting them to an agent.
A checklist for assembling chemistry model workflows
Plan workflows with RDKit, Chemprop, DeepChem and AiZynthFinder by comparing their roles, defining input/output boundaries and separating documented capabilities from integrations that still need evaluation.