Inflammatory bowel diseases, such as Crohn's Disease, are characterised by an altered blood and faecal metabolome, and changes in gut microbiome composition. Here, we present an efficient, scalable, tractable systems biology framework to mechanistically link microbial strains and faecal metabolites. We retrieve strain-level relative abundances from metagenomics data from a cohort of paediatric Crohn's Disease patients with and without dysbiosis and healthy control children and construct and interrogate a personalised microbiome model for each sample. Predicted faecal secretion profiles and strain-level contributions to each metabolite vary broadly between healthy, dysbiotic, and non-dysbiotic microbiomes. The reduced microbial diversity in IBD results in reduced numbers of secreted metabolites, especially in sulfur metabolism. We demonstrate that increased potential to synthesise amino acids is linked to Proteobacteria contributions, in agreement with experimental observations. The established modelling framework yields testable hypotheses that may result in novel therapeutic and dietary interventions targeting the host-gut microbiome-diet axis.