The majority of the discovered structures are crystalline and typically synthesized by hydrothermal processes, [5] but recently it has been discovered that some members of this chemical family can form a stable liquid phase upon heating (prior to thermal decomposition), allowing subsequent supercooling into a glassy state. [6][7][8][9] This work has opened a new glass family, [10] notably distinct from the existing metallic, organic, and inorganic glasses, and with possible applications in, e.g., nuclear waste immobilization, [11] thermoelectrics, [4] gas separation, [12] and energy storage, [13] as well as the possibility to produce bulk-sized hybrid organic-inorganic samples (>1 cm 3 ) without grain boundaries. [14] Although these hybrid glasses are chemically distinct from the traditional glass families, they share a number of structural similarities to chemically ordered network glasses (e.g., oxides), as the metallic nodes (for oxides typically Si, Ge, P; for hybrids typically transition metals) bridge through single oxygens and large organic linkers for oxide and hybrid glasses, respectively. [15] Similarly, the metal nodes in the two glass families share the same range of coordination numbers (typically 4-6). [15][16][17] However, there is a major difference in the discovery Chemical diversification of hybrid organic-inorganic glasses remains limited, especially compared to traditional oxide glasses, for which property tuning is possible through addition of weakly bonded modifier cations. In this work, it is shown that water can depolymerize polyhedra with labile metal-ligand bonds in a cobalt-based coordination network, yielding a series of nonstoichiometric glasses. Calorimetric, spectroscopic, and simulation studies demonstrate that the added water molecules promote the breakage of network bonds and coordination number changes, leading to lower melting and glass transition temperatures. These structural changes modify the physical and chemical properties of the melt-quenched glass, with strong parallels to the "modifier" concept in oxides. It is shown that this approach also applies to other transition metal-based coordination networks, and it will thus enable diversification of hybrid glass chemistry, including nonstoichiometric glass compositions, tuning of properties, and a significant rise in the number of glass-forming hybrid systems by allowing them to melt before thermal decomposition.