Cell-free metabolic engineering (CFME) is emerging as a powerful approach for the production of target molecules and pathway debugging. Unfortunately, high cofactor costs, limited cofactor and energy regeneration, and low volumetric productivities hamper the widespread use and practical implementation of CFME technology. To address these challenges, we have developed a cell-free system that harnesses ensembles of catalytic proteins prepared from crude lysates, or extracts, of cells to fuel highly active heterologous metabolic conversions. As a model pathway, we selected conversion of glucose to 2,3-butanediol (2,3-BD), a medium level commodity chemical with many industrial applications. Specifically, we engineered a single strain of Escherichia coli to express three pathway enzymes necessary to make meso-2,3-BD (m2,3-BD). We then demonstrated that lysates from this strain, with addition of glucose and catalytic amounts of cofactors NAD+ and ATP, can produce m2,3-BD. Endogenous glycolytic enzymes convert glucose to pyruvate, the starting intermediate for m2,3-BD synthesis. Strikingly, with no strain optimization, we observed a maximal synthesis rate of m2,3-BD of 11.3 ± 0.1 g/L/h with a theoretical yield of 71% (0.36 g m2,3-BD/g glucose) in batch reactions. Titers reached 82 ± 8 g/L m2,3-BD in a 30 h fed-batch reaction. Our results highlight the ability for high-level co-factor regeneration in cell-free lysates. Further, they suggest exciting opportunities to use lysate-based systems to rapidly prototype metabolic pathways and carry out molecular transformations when bioconversion yields (g product/L), productivities (g product/L/h), or cellular toxicity limit commercial feasibility of whole-cell fermentation.