Abstract Introduction This study assessed the feasibility and acceptability of FRESH (Families Reporting Every Step to Health), a theory-based child-led family physical activity (PA) intervention delivered online. We also assessed the preliminary effectiveness of the intervention on outcomes of interest and whether pre-specified criteria were met to progress to a full-scale definitive trial. Methods In a three-armed randomised pilot trial, 41 families (with a 7–11-year-old index child) were allocated to a: ‘family’ (FAM), ‘pedometer-only’ (PED), or a no-treatment control (CON) arm. The FAM arm received access to the FRESH website, allowing participants to select step challenges to ‘travel’ to target cities around the world, log their steps, and track progress as families virtually globetrot. FAM and PED arms also received family sets of pedometers. All family members could participate in the evaluation. Physical (e.g., fitness, blood pressure), psychosocial (e.g., social support), behavioural (e.g., objectively-measured PA), and economic (e.g., expenditure for PA) data were collected at baseline, 8- and 52-weeks. Results At 8- and 52-weeks, 98 and 88% of families were retained, respectively. Most children liked participating in the study (> 90%) and thought it was fun (> 80%). Compared to the PED (45%) and CON (39%) arms, a higher percentage of children in the FAM (81%) arm reported doing more activities with their family. Adults agreed that FRESH encouraged their family do more PA and made their family more aware of the amount of PA they do. No notable between-group differences were found for childrens’ minutes in moderate-to-vigorous PA. Sizeable changes of 9.4 (95%CI: 0.4, 18.4) and 15.3 (95%CI: 6.0, 24.5) minutes in moderate-to-vigorous PA was found for adults in the FAM group compared to those in the PED or CON groups, respectively. No other notable differences were found. Conclusion This study demonstrates feasibility and acceptability of the FRESH intervention. All progression criteria were at least partially satisfied. However, we failed to recruit the target sample size and did not find a signal of effectiveness on PA particularly long-term or in children. Further refinements are required to progress to a full-scale trial. Trial registration This study was prospectively registered (ISRCTN12789422) on 16/03/2016.