A robust algorithm for pull-calorimeters that provides a rapid response to changes in respiratory gas exchange has been implemented. Metabolic plateaus (over 20 min), such as that generated by steady treadmill exercise, can be measured accurately (<2.0% error for an energy expenditure level of 16.7 kJ min⁻¹). The time resolution for changes between plateaus can be accurately found with 1 min discrimination. Implementation required only software changes but no structural or instrumentation changes to the chamber. The algorithm was based on the one developed for the push-calorimeter at the Sahlgrenska Hospital in Sweden. The method utilises published equations for the rate of O₂ consumption and CO₂ production in the chamber, along with techniques for suppressing noise and identifying trends. Using the exact solution of the equations for steady state, the O₂ concentrations from the preceding 30 min period are fitted to two connected exponential segments, of variable length, using the least-squares method. The smoothed O₂ concentration and associated time derivative are then determined for the time point 15 min earlier and substituted into the respiration equations. The CO₂ concentrations are subjected to the same analysis. The process is repeated every minute, and the newly computed rates of O₂ consumption and CO₂ production, as well as metabolic rate, are then presented. Gas injection tests proved that the chamber can respond instantaneously to a change from one steady state of respiration to another and correctly averages repeated changes in respiration with periods less than 15 min (<1.4% error for simulated, alternating O₂ consumption levels of 0.81 min⁻¹ and 0.01 min⁻¹). The successful integration of the algorithm into the Pennington chambers allows for traditional 24 h energy expenditure measurements and various metabolic experiments requiring rapid responses.