Matlantis Brings Claude Code Into Atomistic Simulation Platform, Releases Public Skills Library on GitHub
New integration lets researchers create, edit, and run simulations through natural language, extending advanced computational chemistry to non-specialists
CAMBRIDGE, Mass. and TOKYO, May 27, 2026 (GLOBE NEWSWIRE) -- Matlantis, a leading provider of AI-powered atomistic simulation for industrial materials R&D, today announced a new AI agent integration for its universal atomistic simulator that lets researchers build and run simulations using natural-language instructions. The release includes a public Skills library on GitHub, available immediately, and an upcoming installer that will run Anthropic's Claude Code directly inside the Matlantis terminal environment.
Atomistic simulation has long required a combination of computational chemistry knowledge, programming fluency, and environment expertise, a barrier that has kept the technology in the hands of specialists even as the rest of materials R&D has become more cross-functional. Matlantis already removed much of the infrastructure burden by delivering its high-accuracy AI model, PFP (Preferred Potential), as a cloud service. This release targets the last remaining barrier: the scripting layer that sits between a researcher's question and a working simulation.
The integration arrives as AI agents move from experimentation to production use in technical workflows. By embedding Claude Code inside Matlantis and giving it access to a domain-specific Skills library, Matlantis is connecting general-purpose agent capability to the specialized procedures and APIs that simulation work actually requires.
"Simulation has always had three barriers: the infrastructure, the science, and the scripting. We've spent five years removing the first one,” said Daisuke Okanohara, President & CEO, Matlantis. “This release is about removing the third, and it changes who can credibly do atomistic simulation inside an R&D organization."
A public Skills library, available now
The Skills library, released on GitHub, packages Matlantis-specific knowledge—functions, APIs, and representative workflows—into a format that general-purpose AI agents can load and reference. This gives agents access to expertise that isn't present in their underlying training data, allowing them to generate more accurate, context-aware simulation scripts on the first attempt. Initial workflows include structure relaxation, molecular dynamics, reaction pathway exploration, crystal structure prediction, visualization, and retrieving structures from external databases. The library will expand as customer use cases evolve.
Claude Code, embedded in the simulation environment
An installer scheduled for release in a forthcoming update will let users launch Anthropic's Claude Code directly from the Matlantis terminal. Running the agent in-context means researchers can describe what they want to simulate in plain language, generate or edit the underlying scripts, run the calculation, and interpret outputs, all without leaving the workflow. For experimental researchers who lack programming fluency, this opens up simulation as a practical tool. For computational specialists, it removes the repetitive scripting work that typically sits between hypothesis and result.
What it changes for R&D teams
Experimental researchers can now generate and interpret simulations through conversation rather than code. Computational chemistry specialists can compress the time spent reproducing published analyses, screening candidates, and adapting workflows for new systems, adopting a "try first, analyze second" cadence. R&D leaders gain the ability to apply computational methods to a broader range of research themes, including questions that were previously deprioritized because they didn't justify the specialist time required to set them up.
"Matlantis already enables non-specialists to predict material behavior with high accuracy and speed," said Okanohara. "With AI agents now a practical interface for complex tools, we can close the remaining gap between a researcher's question and a running simulation. The goal is to make undiscovered knowledge in the physical world as accessible through natural language as knowledge in the textual world has become."
Availability
The Matlantis Skills library is available now on GitHub. The installer that enables Claude Code to run inside the Matlantis terminal environment will be released in a forthcoming update. For more information, visit matlantis.com.
About Matlantis
Jointly developed by PFN and ENEOS, Matlantis is a universal atomistic simulator that supports large-scale material discovery by reproducing new materials' behavior at an atomic level on the computer. PFN and ENEOS have incorporated a deep learning model into a conventional physical simulator to increase the simulation speed by tens of thousands of times and to support a wide variety of materials. Matlantis was launched in July 2021 as a cloud-based software-as-a-service by Matlantis Corp. (formerly named Preferred Computational Chemistry), a company jointly invested by PFN, ENEOS and Mitsubishi Corporation.
Matlantis is used by over 150 companies and organizations for discovering various materials including catalysts, batteries, semiconductors, alloys, lubricants, ceramics and chemicals. For more information, please visit: https://matlantis.com/en/.
Media Contact:
Emily Townsend
Scratch Marketing + Media for Matlantis
matlantis@scratchmm.com
Legal Disclaimer:
EIN Presswire provides this news content "as is" without warranty of any kind. We do not accept any responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this article. If you have any complaints or copyright issues related to this article, kindly contact the author above.
