¹³C breath testing is increasingly used in physiology and ecology research because of what it reveals about the different fuels that animals oxidize to meet their energetic demands. Here I review the practice of ¹³C breath testing in humans and other animals and describe the impact that contamination by ambient/background CO₂ in the air can have on the accuracy of ¹³C breath measurements. I briefly discuss physical methods to avoid sample contamination as well as the Keeling plot approach that researchers have been using for the past two decades to estimate δ¹³C from breath samples mixed with ambient CO₂. Unfortunately, Keeling plots are not suited for ¹³C breath testing in common situations where (1) a subject's VCO₂ is dynamic, (2) ambient [CO₂] may change, (3) a subject is sensitive to hypercapnia, or (4) in any flow‐through indirect calorimetry system. As such, I present a mathematical solution that addresses these issues by using information about the instantaneous [CO₂] and the δ¹³CO₂ of ambient air as well as the diluted breath sample to back‐calculate the δ¹³CO₂ in the CO₂ exhaled by the animal. I validate this approach by titrating a sample of ¹³C‐enriched gas into an air stream and demonstrate its ability to provide accurate values across a wide range of breath and air mixtures. This approach allows researchers to instantaneously calculate the δ¹³C of exhaled gas of humans or other animals in real time without having to scrub ambient CO₂ or rely on estimated values.