Recent scholarship shows a renewed interest in how political parties connect with various social groups when communicating with voters during election campaigns. However, the scarce available data hinders our ability to gain insights into this aspect of the party-voter relationship. As questions of parties’ appeals to social groups are of high value to our examination and understanding of party behavior via-a-vis their voters, we take a significant step forward to expand available data on group appeals by presenting a new dataset focusing on a cross-national and longitudinal analysis in this paper. Covering 865 general election manifestos of all political parties that gained seats at national-level elections between 1970 and 2024 from Austria, Germany, Ireland, the Netherlands, Spain, Denmark, Switzerland and the UK, this is the first large-scale dataset to allow over time and across-country comparisons of parties’ direct and indirect appeals to social groups. Relying on a comprehensive approach to the concept of group-based appeals developed by Huber and Dolinsky (2023), the dataset is built using advanced computational text analysis methods that combine group and stance detection (positive, negative, neutral) as well as whether policy is included or not. Processing and analyzing parties’ general election manifestos reveal over-time trends in group-based appeals, showcasing within- and between-country similarities and differences in party behavior. This new dataset facilitates a better understanding of group-based appeals via a coherent conceptualization and enhances scholars’ ability to engage in comparative studies.