In the region of Pichilemu (Central Chile), the exhumed roots of the Carboniferous paleo-accretionary wedge developed during subduction of an oceanic realm underneath the western Gondwana margin are exposed. We focus on the areas of Infiernillo beach and Punta de Lobos: the former representing a well-characterized transitional blueschist-facies coherent stack of metabasitic lithologies interlayered at the centimetre-to metre-scale within metasedimentary rocks and glaucophanites, whereas the latter corresponds to a massive, up to hundreds of metres thick, strained metapillow lavas edifice surrounded (not interlayered) by metasedimentary rocks. We provide new field, geochemical, Raman thermometry and zircon SHRIMP U-Th-Pb geochronology on the lithologies forming the Infiernillo-Punta de Lobos complex aiming at characterizing the sedimentary sources and the extent of mechanical mixing. Field observations suggest that the Infiernillo-Punta de Lobos complex can be considered as part of the same successions, subducted and basally accreted almost coherently at high pressure-low temperature (HP-LT) conditions. The geochemistry signal in metasedimentary rocks and maximum deposition ages point to a marked forearc source for all the metasedimentary and glaucophanite lithologies, whereas massive metapillow lavas and greenschists might represent volcanic products and oceanic roughnesses erupted near the trench when the oceanic material approached the Gondwana margin at c. 329 Ma. It is suggested that the interlayering now observed in the field is mostly the consequence of sedimentary and/or volcanic processes close to the trench environment, with minor tectonic reworking of the pre-subduction structures (e.g. lithological contacts). Our results suggest that the Punta de Lobos edifice is closely related to the subduction of large volumes of continent-derived, trench-filling sediments towards HP-LT conditions. Thus, in line with previous experimental and numerical simulations, we emphasize the potential role of oceanic roughness as stress perturbations that could contribute to the subduction of thick sedimentary sequences, and discuss their implications for seismogenesis.