A non-profit research institute

Cures for the conditions no one is studying.

Thousands of rare diseases have no research program, no trial, no one assigned to them. Orphea pairs tireless AI agents with practicing doctors and scientists to investigate the overlooked — funded by the families, friends, and employers closest to each condition, and published openly and free, forever.

7,000+ known rare diseases
~5% have any approved treatment
100% of our findings published open-access
501(c)(3) non-profit · no patents, no paywalls
The gap we exist to close

Some diseases are rare enough that no institution has a reason to study them.

Funding follows scale. A condition affecting a few thousand people worldwide rarely justifies a grant, a trial, or a research team — so the science simply never gets done. The patients are real. The questions are answerable. The work is just unassigned.

95% untreated

The vast majority of rare diseases have no approved therapy — often no active research at all.

4.8yr avg. to diagnosis

Families spend years on a diagnostic odyssey before anyone even names the condition.

<5% of NIH budget

Aggregate rare-disease research is a sliver of funding, scattered across thousands of conditions.

The model · agents + experts

Machine breadth, human judgment.

AI agents do the relentless reading and connecting; clinicians and scientists decide what's worth pursuing and how. Neither works without the other.

PHASE01

Agents survey the unexamined

Autonomous agents continuously read the world's literature, patient registries, case reports, and molecular databases — surfacing neglected conditions and the loose threads no one has had time to pull.

driven by AI agents
PHASE02

Hypotheses, ranked for the overlooked

Agents synthesize cross-disciplinary evidence into testable hypotheses — repurposed drugs, shared pathways, genetic links — then rank them by plausibility, tractability, and how badly they've been ignored.

agents + review
PHASE03

Doctors & scientists decide

A standing board of clinicians and researchers vets every hypothesis, designs the actual study, and partners with the labs and patients who can carry it out. Humans hold the judgment, ethics, and care.

led by experts
PHASE04

Publish everything, immediately

Findings — positive, negative, and inconclusive — are released open-access the moment they're validated, with the data and methods attached so anyone can build on them.

released open

Every published result feeds back into the agents — the more we share, the faster the next discovery.

What your funding builds

You're not donating. You're building a team.

Funding a condition stands up a dedicated unit — practicing experts and AI agents working as one. The further you fund, the more the team can take on: more clinician hours, more agents in parallel, real labs, real trials.

Team assigned to your condition
Active · humans + agents · reporting in the open
The people practicing experts, part-funded by you
Clinical Lead

A practicing physician in the disease area. Sets direction, owns patient care, ethics, and the final call on every study.

Human · principal investigator
Research Scientist

Designs and runs the experiments and analyses the agents propose — turning hypotheses into evidence.

Human · bench & method
Data Lead

Wrangles genomic, molecular, and registry data, and works hand-in-hand with the data agent to find signal.

Human · bioinformatics
Patient Liaison

Keeps the people affected at the center — gathering lived experience and reporting progress back to you.

Human · families & care
The agents AI, working alongside them around the clock
literature-agent

Reads the entire scientific record without stopping and surfaces what's relevant across every field, for the Clinical Lead.

runs 24/7
hypothesis-agent

Connects scattered evidence into testable hypotheses and ranks them by plausibility and how overlooked they are.

proposes & ranks
data-agent

Mines molecular and registry datasets for patterns and candidate targets — paired directly with the Data Lead.

pattern mining
trial-design-agent

Drafts study protocols, simulates outcomes, and flags confounders for the Research Scientist to refine and run.

drafts protocols
The more you fund, the more they can do. A team grows with its support.
Seed
A standing watch
Agents continuously survey your condition; a Clinical Lead reviews what surfaces.
Core
An active program
The full team above, running analyses and publishing findings on your condition.
Full
Labs & trials
Wet-lab partners and clinical studies — the path from hypothesis to a real treatment.
Our binding commitment

We don't own the cure. Everyone does. No patents, no paywalls, no embargoes — just the work, in public.

  • Open-access by default

    Every paper, dataset, and protocol is free to read and reuse the day it's ready.

  • Negative results count

    We publish what didn't work too. A closed door saves the next team years.

  • No commercial capture

    As a non-profit, our only incentive is the science. Discoveries stay in the commons.

  • Reproducible by design

    Methods, code, and agent reasoning ship alongside results — so others can verify and extend.