TLRs play a central role in the innate immune response, recognizing a variety of molecular structures characteristic of pathogens. Although TLR4, together with its co‐receptor myeloid differentiation‐2 (MD‐2), recognize bacterial LPS and therefore Gram‐negative bacterial infections, it also plays a key role in many other pathophysiological processes, including sterile inflammation and viral infection. Specifically, numerous endogenous agonists of TLR4 of notably diverse nature, ranging from proteins to metal ions, have been reported. Direct activation of a single receptor by such a range of molecular signals is very difficult to explain from a structural and mechanistic point of view. It is likely that only a subset of these directly activate the TLR4–MD‐2 complex. We propose three postulates aimed at distinguishing the direct agonists of TLR4 from indirect activators. These postulates are as follows: (i) that the agonist requires the TLR4/MD‐2 receptor complex; (ii) that agonist formed synthetically or in situ must activate the receptor complex in order to eliminate artifacts of contamination by other agonists; and (iii) that a specific molecular interaction between the agonist and TLR4/MD‐2 must be identified. The same type of postulates can be applied to pattern recognition receptors in general.