ChemGraph as an agentic framework for computational chemistry workflows

Summary

ChemGraph is an LLM-driven agentic framework that couples graph-neural-network foundation models with conventional simulation tools to plan, execute, and validate atomistic computational chemistry and materials science workflows from natural-language requests.

Keywords

Cite

@article{2026_pham_chemgraph,
  title = {ChemGraph as an agentic framework for computational chemistry workflows},
  author = {Thang D. Pham and Aditya Tanikanti and Murat Keçeli},
  journal = {Communications Chemistry 9, 33 (2026)},
  year = {2026},
  doi = {10.1038/s42004-025-01776-9}
}