Abstract. We propose that relativistic Galactic jets like those observed in GRS 1915+105 and GRO J1655-40 may produce a small but measurable contribution to the cosmic ray (CR) spectrum. If these jets contain cold protons and heavy ions (as in the case of SS433), it is plausible that this component will consist of a narrow spectral feature, with a mean particle energy corresponding roughly to the bulk kinetic particle energy in the beam, Γ jet m p c 2 . Based on the current estimates of Γ jet , this feature will fall into the range of 3-10 GeV. The presence of several sources with different Γ jet will lead to the superposition of several such peaks. In addition to the narrow peaks, diffusive particle acceleration should also produce a powerlaw, whose low energy cutoff at or above Γ 2 jet m p c 2 would be visible as an additional spectral feature. The large metallicities measured in several binary companions of jet sources suggest that this CR component could have an anomalous composition compared to the bulk Galactic CR spectrum. We provide estimates of the effects of adiabatic losses which are the greatest challenge to models of narrow band CR production in microquasar jets. While the total energy contained in the microquasar CR component is highly uncertain, the local CR spectrum in the vicinity of any microquasar should be severely affected. The upcoming AMS 02 experiment will be able probe the low energy CR spectrum for such components and for composition anomalies. The spectrally peculiar gamma-ray emission produced by interaction of the ISM with CRs surrounding microquasars might be detectable by GLAST. If the presence of a microquasar CR proton component can be ruled out observationally, this argument could be turned around in favor of electron-positron jets. We show that existing OSSE/GRO and future INTEGRAL data on the Galactic 511 keV line flux put interesting constraints on the particle content of microquasar jets. The process of CR production in relativistic flows inside the Galaxy is fundamentally different from the standard picture of CR production in nonrelativistic shocks in supernova remnants, because the particles injected by a relativistic flow are already relativistic, without any need for diffusive acceleration.