Abstract. We present the Brera Multi-scale Wavelet ROSAT HRI source catalogue (BMW-HRI) derived from all ROSAT HRI pointed observations with exposure times longer than 100 s available in the ROSAT public archives. The data were analyzed automatically using a wavelet detection algorithm suited to the detection and characterization of both point-like and extended sources. This algorithm is able to detect and disentangle sources in very crowded fields and/or in the presence of extended or bright sources. Images have been also visually inspected after the analysis to ensure verification. The final catalogue, derived from 4303 observations, consists of 29 089 sources detected with a detection probability of ≥4.2σ. For each source, the primary catalogue entries provide name, position, count rate, flux and extension along with the relative errors. In addition, results of cross-correlations with existing catalogues at different wavelengths (FIRST, IRAS, 2MASS and GSC2) are also reported. Some information is available on the web via the DIANA Interface. As an external check, we compared our catalogue with the previously available ROSHRICAT catalogue (both in its short and long versions) and we were able to recover, for the short version, ∼90% of the entries. We computed the sky coverage of the entire HRI data set by means of simulations. The complete BMW-HRI catalogue provides a sky coverage of 732 deg 2 down to a limiting flux of ∼10 −12 erg s −1 cm −2 and of 10 deg 2 down to ∼10 −14 erg s −1 cm −2 . We were able to compute the cosmological log(N)-log(S ) distribution down to a flux of 1.2 × 10 −14 erg s −1 cm −2 .