<p>Online health resources and large language models are increasingly used as a first point of contact for medical decision-making, yet their reliability in healthcare remains limited by low accuracy, lack of transparency and susceptibility to unverified information. Here we introduce a proof-of-concept conversational self-triage system that guides large language models with 100 medical flowcharts from the American Medical Association, providing a structured and auditable framework for patient decision support. The system leverages a multi-agent framework consisting of a retrieval agent, a decision agent and a conversation agent to identify the most relevant flowchart, interpret patient responses and deliver personalized, patient-friendly recommendations, respectively. Performance was evaluated at scale using synthetic datasets of simulated conversations. The system achieved 84.1% accuracy in flowchart retrieval (<i>N</i> = 2,000) and 99.1% accuracy in flowchart navigation across varied conversational styles and conditions (<i>N</i> = 37,200). By combining the flexibility of free-text interaction with the rigour of standardized clinical protocols, this approach demonstrates the feasibility of transparent, accurate and generalizable AI-assisted self-triage, with potential to support informed patient decision-making outside clinical settings while improving healthcare resource utilization.</p>