Despite the wide synthetic potential of diazo compounds (XÀH insertion, ylide formation, cyclopropanation, cycloaddition etc.), [1] concerns over the hazards associated with their preparation, isolation, and use have hindered their full exploitation in both academic and industrial laboratories. A few diazo compounds are commercially available (e.g. ethyl and butyl diazoacetate, TMS-diazomethane and diazodimedone), [2] but safe and convenient access to a wider range of useful functionalized diazo species is still desirable.[3] Diazo transfer can be used to access a-diazocarbonyls, but this only partially addresses the safety concerns associated with the diazo species, as the use of equally hazardous azidebased diazo-transfer reagents is still required.[4] Ideally, it would be beneficial if the diazo species could be generated and consumed in situ so that handling of the hazardous diazo compound is avoided altogether.Recent work by Ley, [5] Jamison, [6] Kappe, [7] and others [4,[8][9][10] has shown that highly reactive diazo and azido compounds can be used in lab-scale continuous-flow reactors to achieve a number of very useful synthetic transformations, and indeed work from our own laboratory has shown that ethyl diazoacetate can be used in-flow to access b-keto esters.[11] We therefore wondered if it was possible to actually generate a-diazocarbonyl compounds under flow conditions and then use these materials directly in further synthetic manipulations, thus minimizing exposure to any potentially hazardous material. In effect, could we develop a continuous-flow diazo generator and then demonstrate its use to prepare a range of useful a-alkoxy (3 a-i) and aamino acid (4 a-i) derivatives through O À H and N À H insertion (Scheme 1)?At the outset we were aware that in order to provide an acceptable solution to the problem, we needed to identify a way to access the diazo compounds of interest (2 a-i) from starting materials that showed an acceptable safety profile, that is, the precursor molecules and reagents should be safer to prepare and handle than the diazocarbonyl compounds being produced. Of the methods available for the generation of a-diazocarbonyl compounds, we were particularly attracted to the Bamford-Stevens reaction as it uses readily accessible arylsulfonylhydrazones (e.g., 1 a-i) as starting materials, with the corresponding diazocarbonyls being generated upon exposure to relatively weak base at moderate reaction temperatures. [12][13][14] Thermal stability studies (DSC and TGA) were conducted on the tosylhydrazone 1 b and its corresponding methyl diazoester 2 b [15,16] in order to determine if a safe window of operation could be identified for the continuous-flow process (see the Supporting Information). The results clearly show that the rate of initial mass loss from diazoester 2 b peaks at 125 8C, which corresponds to a significant exotherm. In comparison tosylhydrazone 1 b has a rate of mass loss which peaks at 221 8C, indicating that it is substantially more thermally stable.[15] We therefore co...