One trusted place for rare disease information, education, and health records, built so every answer traces back to a vetted source.
Built by Freebranch for RareStrides
People living with a rare disease often search for years, and the information they find is scattered across journals, forums, and websites of uneven quality. Clinicians face the same problem from the other side: it is hard to know which sources to trust, and harder still to keep up.
CodeRare, one part of RareStrides, brings patients and caregivers, clinicians, and rare disease organizations onto one vetted record. Every fact carries a citation and a credibility score, nurses review content before it goes live, and an AI assistant named Z answers questions only from that reviewed content.
Around that knowledge base sit continuing education courses, an organization directory, and a health record that patients control. Sign-up is by invitation while the platform grows.
A knowledge base, a review process, and the tools each community needs, all working from the same vetted record.
Each profile is organized into Rare Model of Care sections, from symptoms and genetics to tests and resources. Diseases can carry several ICD-10 codes, and every fact links to its source.
Publishers earn reputation tiers and evidence is graded with the GRADE framework. The combined CodeRare Score only changes when enough reviewers approve, and every change is recorded in an audit trail.
CME modules built from video, slide, and quiz lessons, with badges for learners. An organization approves each course, CodeRare vets it, and that vetting must be renewed every six months.
Patients keep their own record and decide who sees it. Sharing is scoped, time-limited, and revocable, every read is logged, and administrators get no automatic access.
Rare disease organizations get a place in the directory and a self-service portal to manage their profile and contribute content for review.
Z routes each question to specialist agents for diseases, anatomy, courses, and the signed-in patient's own records, then answers with citations to the records it used.
CodeRare uses AI in two places: the Z assistant, and a research pipeline that drafts new content. Both follow the same rule. People decide what counts as knowledge.
Z's specialist agents skip unreviewed AI drafts and any fact below the credibility score an administrator sets. When they find nothing, Z says so instead of guessing.
Drafts from the AI research pipeline land in a nurse vetting queue. Nothing the pipeline writes goes live until a person has reviewed it.
Any diagnosis code the AI suggests is checked against the real ICD-10 table, and codes that don't match are dropped. An unreviewed AI suggestion can never become a disease's primary code.
Z can read health records only for the signed-in patient, and turns away anyone who isn't signed in. Sharing grants are scoped and revocable, and every read is logged.
Every project here started with a specific problem. Tell us about yours, and we'll give you an honest read on what a custom build would take.