The LivinGut project explores the bi-directional relationship between gut microbiome states and social norms related to food behaviors, particularly fermented products. By integrating Living Labs, we will assess how microbiome composition influences food perception and how dietary norms impact microbiome stability. This proof-of-concept study will compare two microbiome groups: diverse and rich, low-diversity and Bacteroides-dominant. Through AI-driven text mining of discussions, we will identify barriers and levers to fermented food consumption. Structured in a DBTL cycle, the project includes four work packages. First, we will classify 72 participants from the French Gut cohort into microbiome-based consumer segments. Next, Living Labs will generate a 150,000–200,000-word corpus, providing insights into sensory biases and perception gaps. AI-powered large language models (LLMs) will then analyze these discussions, detecting emotional responses, lifestyle factors, and microbiome-linked narratives. We will adapt an open-source LLM and integrate AI tools with SIDURI’s data warehouse. Finally, validated insights will refine consumer perception surveys and inform industry strategies for fermented food positioning. By bridging microbiology, consumer science, AI, and ethics, LivinGut will produce actionable knowledge for food stakeholders, advancing personalized nutrition and gut microbiome research