This study examines how political virality in Indonesia’s digital public sphere was produced and experienced through the interaction of deepfakes, algorithmic mediation, and platform power during the national political protests of 25 August 2025. Rather than understanding political virality solely as the rapid circulation of content, this study conceptualizes it as a communicative experience shaped by algorithmic logics, collective affect, and power relations within platformized media environments. Drawing on a digital phenomenology approach, the research explores users’ subjective experiences in encountering and interpreting viral political content on social media. Data were collected through in-depth interviews, digital observation, and media artifact analysis related to the circulation of deepfakes and other viral political content during the protests. The data were analyzed interpretively to identify recurring patterns of perception, emotional response, and digital communication practices. The findings reveal that deepfakes function not merely as forms of visual manipulation but as mechanisms for producing political reality, influencing public perception, emotional engagement, and the legitimacy of information. Simultaneously, platform algorithms amplify emotionally charged content through visibility-enhancing mechanisms, fostering virality-driven political participation. Four key dynamics emerged from the analysis: (1) digital political experiences are characterized by epistemic uncertainty and heightened emotional intensity; (2) algorithms operate as curatorial agents that regulate content visibility and circulation; (3) perceptions of power are shaped more by virality than by factual verification; and (4) subjective political experiences are constituted through the interaction of manipulative content, algorithmic mediation, and digital power relations. The study contributes to critical digital communication scholarship by advancing understanding of how political reality, visibility, and power are reconfigured within contemporary platform cultures and AI-mediated public spheres.