Amsterdam Team Builds $5,000 Open-Source Chemistry Robot
Get the Health newsletter
Daily health & science — research, biotech, public health, the studies worth knowing. Free.
- Prof. Timothy Noël's group at the University of Amsterdam's Van 't Hoff Institute for Molecular Sciences published RoboChem Flex in Nature Synthesis, cutting the system's estimated cost to about $5,000 from the original RoboChem's $50,000-plus price tag.
- RoboChem Flex uses readily available or 3D-printed components and is controlled by the open-source OmniPlatypus software package, with all code, firmware, and hardware schematics available on GitHub so any lab can build its own copy.
- The system integrates a modular Bayesian Optimization AI agent and supports plug-and-play connections to inline analytical instruments including NMR, UHPLC-MS, and Raman spectroscopy for fully autonomous 24/7 reaction optimization.
- For labs that can't afford full inline analytics, the team added a low-cost 3D-printed liquid sampling unit enabling a 'human-in-the-loop' workflow using shared existing analytical equipment, keeping entry costs close to the $5,000 baseline.
- The Noel group demonstrated RoboChem Flex across six case studies spanning photocatalysis, biocatalysis, and thermal cross-coupling, and separately validated the proposed syntheses by performing them manually in their lab.
- The original RoboChem, published in Science in early 2024, could autonomously optimize synthesis of 10 to 20 molecules — work that would take a PhD student several months — and Flex is designed to extend that capability to resource-limited groups.
Why it matters: By slashing the price of an AI-driven autonomous synthesis platform roughly tenfold to $5,000 and publishing every design file, the Noel group effectively removes the capital barrier that has confined self-driving lab technology to well-funded institutions, letting smaller research groups run machine-learning-guided reaction optimization on the same problems in photochemistry, biocatalysis, and cross-coupling that were previously off-limits.
