<p>Over the past few decades, YouTube has become a central arena for public narratives in which global risks, such as Climate Change, are increasingly framed for online audiences. We present empirical research conducted through a novel Framing Analysis approach, which we define as “Frame Sequencing”, aimed at analysing large volumes of unstructured data, particularly valuable for qualitative and quantitative inquiries on audiovisual content online. Methodologically, we offer a replicable workflow for a multi-layer Framing Analysis of video data, showing how context evolution shapes the prominence and articulation of communication practices. Our Frame Sequencing approach uses video captions as a first proxy to analyse video content, while annual splitting, vectorisation, and hierarchical clustering produce the first output: the Textual Frame Sequences (TFS), a matrix that arranges the twenty most frequent terms per cluster across the yearly snapshots. From each cluster we selected three Prototype videos which are then qualitatively analysed, thus producing the second output, the Visual Frame Sequences (VFS): a matrix juxtaposing the extracted frames of the prototypes. Finally, we tested the proposed approach on the YouTube platform following the discussed issue of Climate Change, where it insightfully highlighted its frame sequences on which a qualitative deepening has been carried out, leading to the acknowledgement of three evolving narratives portrayed by influential platform actors: “Data room narratives”, “Science and call to action”, and “Climate Change political debate”.</p>