FlynnMed brings structured health records, deterministic safety checks, evidence retrieval and role-aware clinical support into one traceable workspace for patients, carers and healthcare professionals. Explore FlynnMed
Combines symptoms, medicines, allergies, observations, documents, history and patient preferences into a structured longitudinal record.
Applies consent and access controls, deterministic safety rules, clinical policy checks, and trusted evidence retrieval before generating role-aware clinical support.
Connects patients and healthcare professionals through consented information exchange, evidence-backed answers, care plans, pre-visit summaries and clinical review, with audit and evidence lineage maintained
A personal health workspace that helps patients and carers organise their health information, understand supported risks and stay actively involved in their care.
Key features
A consent-controlled clinical workspace for healthcare professionals to review patient context, access evidence and prepare clinical actions within their professional role.
Key features
FlynnMed is built as an inspectable open-source clinical software stack using React and TypeScript, FastAPI, PostgreSQL and modular Python services. The safety, evidence, policy and orchestration layers can be independently reviewed, tested and extended, with the platform released under the MIT Licence.
FlynnMed includes an optional MCP server that exposes selected clinical capabilities to approved AI clients and agents through controlled tool interfaces. These tools support functions such as patient-context retrieval, clinical-output validation, note generation, trial search and email delivery, with authenticated HTTP access available through the FastAPI application.
Clinical workflows pass through deterministic safety rules and policy controls before supported information is presented or acted upon. Current checks cover defined high-risk scenarios including potassium abnormalities, emergency symptoms, medicine–allergy conflicts and selected warfarin–NSAID combinations, while policy gates constrain responses according to role, context and permitted workflow.
The retrieval layer searches authoritative clinical guidance, medicines data, biomedical literature and clinical-study services, including NHS/NICE, CDC, MedlinePlus, openFDA, Europe PMC/PMC and ClinicalTrials.gov. Source provider, jurisdiction and available reuse metadata are retained so evidence can be ranked and presented without treating guidance from different jurisdictions as interchangeable.
Retrieved evidence is extracted, ranked and evaluated before being incorporated into a clinical response. FlynnMed preserves source provenance, handles conflicting evidence, links claims back to supporting passages and applies policy checks before generating a role-appropriate answer, allowing important claims to remain traceable to their underlying evidence.
FlynnMed records the relationship between patient context, retrieved evidence, generated claims and clinical workflow events. Consent records, audit events, evidence ledgers and answer-claim lineage provide a reviewable history of how information was accessed, what evidence supported an output and how the resulting clinical interaction progressed.