Autonomous laboratories can accelerate discoveries in chemical synthesis, but this requires automated measurements coupled with reliable decision-making^1,2. Most autonomous laboratories involve bespoke automated equipment^3,4,5,6, and reaction outcomes are often assessed using a single, hard-wired characterization technique⁷. Any decision-making algorithms⁸ must then operate using this narrow range of characterization data^9,10. By contrast, manual experiments tend to draw on a wider range of instruments to characterize reaction products, and decisions are rarely taken based on one measurement alone. Here we show that a synthesis laboratory can be integrated into an autonomous laboratory by using mobile robots^11,12,13 that operate equipment and make decisions in a human-like way. Our modular workflow combines mobile robots, an automated synthesis platform, a liquid chromatography–mass spectrometer and a benchtop nuclear magnetic resonance spectrometer. This allows robots to share existing laboratory equipment with human researchers without monopolizing it or requiring extensive redesign. A heuristic decision-maker processes the orthogonal measurement data, selecting successful reactions to take forward and automatically checking the reproducibility of any screening hits. We exemplify this approach in the three areas of structural diversification chemistry, supramolecular host–guest chemistry and photochemical synthesis. This strategy is particularly suited to exploratory chemistry that can yield multiple potential products, as for supramolecular assemblies, where we also extend the method to an autonomous function assay by evaluating host–guest binding properties.

自主实验室可以加速化学合成的发现，但这需要自动化测量和可靠的决策^1,2。大多数自主实验室都涉及定制的自动化设备^{3,4,5,6，并且}通常使用单一的硬连线表征技术⁷ 来评估反应结果。然后，任何决策算法⁸（decision-making algorithms）都必须使用这个狭窄的特征数据^范围9,10来运行。相比之下，手动实验往往使用更广泛的仪器来表征反应产物，并且很少仅根据一次测量做出决定。在这里，我们展示了通过使用移动机器人^11,12,13 可以集成到自主实验室中，这些机器人以类似人类的方式操作设备并做出决策。我们的模块化工作流程结合了移动机器人、自动合成平台、液相色谱-质谱仪和台式核磁共振波谱仪。这使得机器人能够与人类研究人员共享现有的实验室设备，而不会垄断它或需要大量重新设计。启发式决策者处理正交测量数据，选择成功的反应进行推进，并自动检查任何筛选结果的可重复性。我们在结构多样化化学、超分子主客体化学和光化学合成三个领域中举例说明了这种方法。这种策略特别适用于可以产生多种潜在产物的探索性化学，例如超分子组装，我们还通过评估主客体结合特性将该方法扩展到自主功能测定。