Digitalization in materials science is essential for transparent and efficient research. Laboratories require a digital infrastructure for their research processes to accumulate, reuse, and plan experiments based on structured data. Large-scale projects involving multiple institutions need flexible platforms to integrate heterogeneous experimental and computational data.

We present MatInf, a customizable Research Data Management System (RDMS) designed to support diverse data types and workflows in materials science. The system is based on extensible data types built from a core simple object, which includes basic metadata. Specialized types – such as chemical system, chemical compound, literature publication, and handover – can be further extended to define domain-specific entities like experiment plan, nano-powder sample, composition measurement, and others.

Each type is defined by templates that specify the structure of data and metadata, either as key-value pairs or structured tables. This enables accurate modeling of real-world research entities. Objects can be associated with documents, and the system provides a flexible API for validating, extracting, and visualizing object content via external services. This allows seamless integration of custom data formats and automation of metadata extraction.

MatInf supports user-defined project trees and directed typed relationships between objects, enabling construction of a Directed Weighted Multigraph – a foundation for domain-specific Knowledge Graphs. The platform includes a user-friendly interface for advanced search and secure access, along with an API for custom SQL-based queries.

MatInf is multi-tenant, scalable, and in active use within major materials science initiatives, including CRC 247, CRC 1625, and ERC DEMI.