About Constellate
Your family is not the first. What you need may already exist under another disease's name.
Families with a rare diagnosis and the groups that support them often work alone, because the registries, studies and tools they need sit under a neighbouring disease with a different name. Constellate is an evidence-graded map of rare diseases that finds those connections and shows the proof for every one of them.
One mother, two moments
A composite figure for the story, not a real person.
Moment 1
We just got the diagnosis
A family has a name for a rare disease and nobody to ask. Constellate shows them the patient groups for that diagnosis, the studies that list it and, when there is nothing, exactly what was searched.
Moment 2
I lead a patient group
Months later the same mother leads a small group. Constellate shows related communities, and the registries, natural history studies and outcome scales that already exist, so she joins instead of building from zero.
What the app does
Start with the diagnosis
Search by disease, gene or symptom, in any of its names.
Find the related community and the asset it already has
Diseases are linked by shared genes and phenotypes. Each line shows its source, date, evidence level and counter-evidence.
Get an action card
What transfers, what differs, what needs expert review. Every sentence cites the evidence it rests on, and the card can be exported.
Honest gap
When there is no trail, we say so
Where no source supports a link, Constellate does not fill the space. It shows what was searched and when, which communities are nearest, and what would have to be tested next. A missing answer is a result, and it is marked as one.
Team
- Hanno Kügler · tech lead, pipeline and app
- Osman · operations and curation
- Recep · story, evidence and video
Built at the Hack-Nation 7th Global AI Hackathon, Vienna hub, challenge C5 “AI Atlas for the World's Rare Diseases”.
AI inside
Language models extract claims from abstracts, match organisation names to diseases and explain an action card in plain language. The pipeline is built on the OpenAI API (Extract, Reconcile, Explain); the current data export ran these steps on Claude (Anthropic), and every evidence item names its model. Each use is marked in the product, and every explanation sentence must cite an evidence edge. Details on How we know.
A research-navigation tool, not medical advice. Decisions about care and study participation belong with a family's clinicians and the study team.