We consider the Kuramoto model of an ensemble of interacting oscillators allowing for an arbitrary distribution of frequencies and coupling strengths. We define a family of traveling wave states as stationary in a rotating frame, and derive general equations for their parameters. We suggest empirical stability conditions which, for the case of incoherence, become exact. In addition to making new theoretical predictions, we show that many earlier results follow naturally from our general framework. The results are applicable in scientific contexts ranging from physics to biology.Almost every real-life physical system involves a large number of interacting subsystems. In many cases, they can be treated as a population of interacting phase oscillators that can be described in terms of the Kuramoto model (KM) [1], which we consider in the forṁHere θ i , ω i and K i are respectively the ith oscillator's phase, natural frequency, and strength of coupling to the other oscillators; and ω i and K i are randomly chosen from a probability density g(ω, K). The KM has been used in a variety of applications, ranging from brain dynamics and human crowd behavior to Josephson junction arrays and neutrino flavor oscillations [2][3][4][5], so that the analysis of its dynamics is of high topical interest and broad applicability in science. Many KM modifications have been considered, e.g. with: nonisochronicity [6][7][8] However, the basic model (1) remains generally unsolved. Although the recent OA-ansatz [17][18][19] provided an important advance, a full reduction of the dynamics of (1) to a set of ODE is possible only in the case of multimodal-δ K and multimodal Lorenzian ω distributions; and the complexity of the equations obtained grows with increasing multimodality for either variable. Thus KM solutions have been obtained only for particular cases of g(ω, K), e.g. constant K for a frequency distribution that is unimodal and symmetric (classic KM [1]) or bimodal-Lorenzian [20,21]; or bimodal-δ distribution of K with unimodal Lorenzian ω [22], etc. But no attempt has been made to solve (1) in general. In this Letter we develop a framework to treat (1) for arbitrary g(ω, K), thus taking a major step towards filling this gap.We start with basic definitions. The collective behavior of KM oscillators is described by the order parameterwhere R is the strength of the mean field created by all oscillators, quantifying the "agreement" between them.It is usually the main quantity of interest. In the continuum limit, N → ∞, (1) is treated using the probability density function (PDF) f (θ, ω, K, t), i.e. the probability that an oscillator has phase θ, coupling strength K and frequency ω at time t. Usually [18,19,23,24] the PDF can be represented by the OA ansatz [17]where α = α(ω, K, t) should satisfyand the integrals are taken over (−∞, ∞) if unspecified. The KM equations (1) do not change form when transformed to a frame rotating at Ω (θ i → θ i − Ωt): this is equivalent to changing the frequency distribution g(ω, K) → g(ω + Ω, K), where...