In this paper we describe and publicly release a catalogue consisting of 8445 point-like X-ray sources detected in the XMM-XXL north survey. For the 2512 AGN which have reliable spectroscopy from SDSS-III/BOSS, we present the X-ray spectral fitting which has been computed with a Bayesian approach. We have also applied an X-ray spectral stacking method to different sub-samples, selected on the basis of the AGN physical properties (L 2−10 keV , z, M BH , λ Edd and N H ). We confirm the well-known Iwasawa-Taniguchi effect in our luminosityredshift sub-samples, and argue that such an effect is due to a decrease in the covering factor of a distant obscuring 'torus' with increasing X-ray luminosity. By comparing the distribution of the reflection fraction, the ratio of the normalization of the reflected component to the direct radiation, we find that the low-luminosity, low-redshift sub-sample had systematically higher reflection fraction values than the high-redshift, high-luminosity one. On the other hand, no significant difference is found between samples having similar luminosity but different redshift, suggesting that the structure of the torus does not evolve strongly with redshift. Contrary to previous works, we do not find evidence for an increasing photon index at high Eddington ratio. This may be an indication that the structure of the accretion disc changes as the Eddington ratio approaches unity. Comparing our X-ray spectral analysis results with the optical spectral classification, we find that ∼ 20 per cent of optical type-1 AGN show an X-ray absorbing column density higher than 10 21.5 cm −2 , and about 50 per cent of type-2 AGN have an X-ray absorbing column density less than 10 21.5 cm −2 . We suggest that the excess X-ray absorption shown in the high-luminosity optical type-1 AGN can be due to small-scale dustfree gas within (or close to) the broad line region, while in the low-luminosity ones it can be due to a clumpy torus with a large covering factor.