Cellobiose, a β-(1 → 4)-linked disaccharide indigestible by humans, is a novel candidate prebiotic. Evidence from controlled human trials on its effects on the gut microbiota and metabolites remains limited. We conducted an exploratory randomised, double-blind, placebo-controlled trial in 37 healthy adults. Participants were allocated to three arms for 4 weeks, received 5 g/day and 10 g/day for cellobiose, oligofructose P95, or maltodextrin placebo. The primary endpoint was change in absolute abundance of Bifidobacterium and Lactobacillus (measured by 16S rRNA gene sequencing). Secondary endpoints included other taxa, diversity, faecal short-chain fatty acids (SCFAs) measured by gas chromatography–mass spectrometry and gastrointestinal (GI) tolerability. Cellobiose did not significantly change Bifidobacterium or Lactobacillus versus baseline or placebo (P > 0.05). Oligofructose P95 induced a borderline increase in Bifidobacterium (P = 0.049). Overall community composition remained unchanged. However, network analyses under cellobiose revealed tighter positive correlations among key genera. Faecal SCFA levels were not altered. GI symptom rates were low and similar across all arms. Cellobiose, at doses up to 10 g/day, is safe and well tolerated but did not enrich classic beneficial taxa. The findings suggest cellobiose drives subtle community shifts towards butyrate-producing bacteria.