You, checked against every trial.

• Identifying details are removed on our server — names, dates of birth, record numbers, addresses — before anything is read.
• The documents are never saved. What we keep is a summary of the diagnosis and treatment, in a form with no field for a name.
• We never ask who you are. No account, no email. Your search is tied to a code in your browser and nothing else — which is why, if you lose it, we cannot look you up.

Clinic letters, pathology, genomic testing, discharge summaries. PDF or Word. They need not be complete or tidy — we say what is missing.

Each trial lists who can join and who cannot — often fifty lines. We check your record against every line and mark it met, not met, or not stated.

It will not tell you whether a trial is right for you. That is a conversation with the doctor treating you — your oncologist, neurologist, rheumatologist, whoever that is — and this is meant to make it shorter and better informed, not to replace it.

A trial is only ruled out when a criterion clearly excludes you. Being cautious that way means some of what we surface will not pan out. We would rather show you a trial you turn out not to fit than hide one you did.

Matching runs over the full ClinicalTrials.gov corpus, with eligibility criteria parsed per study rather than searched as free text. Retrieval combines MeSH concept expansion, Postgres full-text and vector similarity. Screening assesses each criterion individually and records met, not met or unknown, with the evidence used.

A trial is excluded only on an explicit exclusion match. Inclusion criteria the record contradicts are surfaced on the trial rather than used to filter, since basket trials contradict most of their own cohort criteria by design.

We reload the registry every week. Pay once and we re-run your search against it for 90 days, and email you when something new fits — never anything about your diagnosis, just that there is something to see.