Substantial reductions in meat consumption are crucial to attain climate targets and limit the numerous negative impacts associated with livestock farming. This study empirically tests the effectiveness of different policy interventions to reduce the carbon footprint of food choices using an incentive-compatible online randomised controlled trial with 4000 participants. The experiment uses an interactive web-platform that mimics popular online takeaway platforms (such as UberEats), and includes three treatment conditions: a meat tax, a carbon footprint label, and a choice architecture intervention, as well as a business-as-usual control condition. The meat tax treatment imposes a surcharge on meat dishes, while the label treatment provides information on the carbon footprint of each meal. The choice architecture intervention re-orders the menu so that the lowest carbon-impact restaurants and dishes are presented first. Exploratory analysis will also provide insights into potential hedonic welfare effects associated with the interventions, as well as whether experience with the interventions has an effect on acceptability of sustainable food policy interventions. Findings will provide valuable insights into the relative efficacy of different types of interventions (both conventional and behavioural interventions) tested in the same experimental setting. An initial pilot with 200 participants was successfully completed and data collection for the main experiment will be completed by March 2023. Our study is one of the first studies to specifically focus on a meat tax, which is subject to much debate, however only few studies have attempted to experimentally assess its potential impact on behaviour. Moreover, our study examines a novel and increasingly relevant food choice setting, specifically, the market for online takeaways, which is predicted to experience substantial growth in the near future.