The NINFEA cohort is an Italian birth cohort aiming at recruiting pregnant women through the Internet and following-up their children. To understand whether Internet-based recruitment was feasible we started a pilot in July 2005 targeted to pregnant women visiting the hospitals of the city of Turin (900,000 inhabitants), where we advertised the study. For this purpose we constructed a website (www.progettoninfea.it), with on-line questionnaires to be completed during pregnancy and at 6 and 18 months after delivery. Participants’ characteristics were compared with those of women giving birth in Turin, which are routinely released as individual anonymous records. We also compared complete with partial respondents. We also carried out a survey of 122 women giving birth in the main Turin obstetric hospital to estimate the proportion of pregnant women with access to the Internet and awareness of the NINFEA cohort. By December 2006, we had recruited 670 women. Participation was associated with being older, a university graduate, primiparous and born in Italy. Complete response (n = 633) was associated with being primiparous and participation after the first trimester. In the survey, 66% (95% confidence interval: 56–74%; n = 80) of the women had access to the Internet and 42% (33–51%; n = 51) were aware of the study; 6.5% (2.9–12.5%; n = 8) had participated in the NINFEA cohort. Our study indicates that recruitment of an Internet-based birth cohort is feasible. As with many other types of birth cohort study, the participants are a self-selected sample. To minimise selection bias analyses should therefore be limited to internal comparisons.