VitriFlex: An Open-Source, Modular, and Customizable Robotic Platform for Cryo-EM Grid Preparation

Specimen preparation remains the principal bottleneck in cryo-electron microscopy (cryo-EM). Achieving uniformly thin ice without denaturation, preferred particle orientation, or air-water-interface damage is challenging. The demands of time-resolved experiments and sub-nanoliter sample volumes raise the bar even further. Commercial plungers are closed, high-cost platforms that limit experimental flexibility, constraining users to a few vendor-defined "custom" routines. By contrast, an open, modular and automated system would allow researchers to tailor experimental conditions by tweaking mixing, dispensing, and freezing parameters while lowering barriers to adoption. We present VitriFlex, an open-source cryo-EM grid preparation platform built around an industrial-grade SCARA robot and 3D-printed components. Designed for modularity and ease of assembly, the system enables programmable control of grid handling, sample deposition, and blotting, supporting both standard and time-resolved workflows. VitriFlex's modularity allows flexible implementation of sample delivery strategies, including pre-mixing and on-grid mixing modes, combined with acoustic-assisted spray application. The system allows for spray-to-plunge delays as short as ~130 ms, enabling initiation of biochemical reactions within sub-second timescales prior to vitrification. Protocols are easily modifiable, and new components can be integrated with minimal redesign. We validated the platform using standard single-particle samples including apoferritin and dGTPase. Additionally, we demonstrate effective mixing capabilities by generating spike-ACE2 and α7-bungarotoxin complexes. In all cases, VitriFlex reproducibly yielded high-quality, collectable grids, allowing us to achieve high-resolution reconstructions. These results highlight VitriFlex as a flexible, accessible solution for both routine and time-resolved cryo-EM applications.