Understanding cell-cell communication (CCC) pathways from single-cell or spatial transcriptomic data is key to unraveling biological processes. Recently, multiple CCC methods have been developed but primarily focus on refining ligand-receptor (L-R) interaction scores. A critical gap for a more comprehensive picture of cellular crosstalks lies in the integration of upstream and downstream intracellular pathways in the sender and receiver cells. We report here CausalCCC, https://miic.curie.fr/causalCCC.php, an interactive web server, which addresses this need by reconstructing gene-gene interaction pathways across two or more interacting cell types from single-cell or spatial transcriptomic data. CausalCCC includes a graphical introduction and a demo dataset within the workbench page as well as a comprehensive tutorial. CausalCCC methodology integrates a robust and scalable causal network reconstruction method, multivariate information-based inductive causation, with internally computed L-R pairs using LIANA+ (including CellphoneDBv5, SingleCellSignalR, Connectome, NATMI, and Log2FC). Alternatively, user-defined L-R pairs from any CCC methods can also be uploaded. We showcase here CausalCCC on different single-cell and spatial transcriptomic datasets from three original CCC methods (NicheNet, CellChat, and Misty). CausalCCC web server offers unique interactive visualization tools dedicated to single-cell data practitioners seeking to go beyond L-R scores and explore extended CCC pathways across multiple interacting cell types.