The present research work caters to two important needs of rural India: i) desalination of subsoil/coastal brackish water and ii) basic refrigeration for short term preservations of agro-produce, medicines etc. Fortunately, such places are blessed with abundant solar insolation and/or low grade thermal energy (< 100°C) is available which may be tapped for this purpose instead of relying solely on grid electricity. Both the objectives of desalination and cooling are realized by evaporating brackish water at a low pressure (~1 kPa) and thermally compressing the water vapour to a higher pressure before condensing it. Adsorption route is chosen for compression where silica gel is the adsorbent and water to be desalinated as the refrigerant. The objective of this study is to develop a laboratory prototype of a two-stage adsorption cooling cum desalination system driven by low grade heat source. The entire system is air-cooled which is necessitated by non-availability of heat exchange grade cooling water. Initially various experimental and theoretical studies are carried out for characterizing silica gel + water pair which is fundamental to the system design. RD type silica gel is used in this study due to its high uptake capabilities. The uptakes for this adsorption pair at various pressure and temperature conditions are measured using a specially designed isothermal adsorber cell connected to an evaporator. Subsequently, a modelling study of adsorption kinetics is performed for a monolayer of silica gel in order to estimate the adsorption time scale. This time scale is used as an input for the scaling analysis of columnar packed silica gel bed. The scaling analysis showed that the thermal diffusion time scale limits the adsorption process. It also showed that for a given thermal length scale, the bed has a unique vapour flow length scale beyond which the adsorption phenomenon gets limited due to pressure drop. The scaling results are validated by simulation studies. A shell-and-tube heat exchanger is chosen for the adsorber which closely mimics the columnar silica gel packing studied in scaling analysis. The heat exchanger is designed for radial entry of vapour. A modelling study is performed on ANSYS® Fluent platform for optimising the tube pitch by minimising the overall thermal capacitance of the bed. The shell diameter is determined for this tube pitch based on the vapour flow length criterion established through scaling. To experimentally study the effect of pressure drop on bed performance, the radial entry of vapour is closed for 1 bed/stage (out of the 4 beds/stage) enforcing the vapour to flow along the longer axial dimension. The system is generously instrumented for precise measurements and control over the various experimental parameters. For the functioning of the adsorber system, various vapour valves and water (heat transfer fluid) valves need to be operated in a cyclical and synchronized manner. Individual components are fitted with pressure, temperature and water flow sensors. The entire operation and data acquisition for the adsorption system has been automated using National Instruments® (NI) PXIe controller executing an in-house developed code written on NI Labview® platform. To simulate solar/waste heat input, multiple electrical heaters are used in this study and a constant temperature bath is used to simulate the cooling load at the evaporator. Prior to conducting experiments a 4-bed lumped dynamic model is developed based on the design data of the system to simulate the two-stage system performance for various input conditions. The study helped to optimise the performance of a two-stage system. The study also compares the two-stage and single-stage system performance for various ambient temperatures (25–40°C). The study revealed that for pressure lifts higher than 5 kPa a two-stage system is preferable. A detailed experimental study is conducted on the developed prototype by operating it in various conditions namely 2, 3 and 4–bed modes for single and two-stage operations; with 1.0–1.7 kPa evaporator pressures, half cycle time varying between 1200–3000s and source temperature in the range of 75–85°C. The system is operated indoors during summer conditions wherein the ambient temperature is found to be 36±1°C which is significantly higher than the design point of 25°C. This resulted in lower than expected throughput; however, the system performance variation is qualitatively similar to as predicted by the lumped model. A comparison between the experimental and simulated bed temperature revealed that the thermal wave during the switching of hot/cold water plays a significant role causing a large deviation from the simulation results. A comparative study is carried out between the beds with radial vapour flow to that with axial flow and the results validate the scaling criterion. Experimental results also depict that two-stage operation is favourable when the pressure lift required is larger than 5 kPa. Such large pressure lift is encountered when air-cooling is used in a tropical country like India.