About the data
TrialPath is a search and summarization layer over ClinicalTrials.gov. We don't own, host, or modify the underlying data — every result you see comes from the public API, fetched fresh and cached for at most an hour.
The source
All trial data comes from the ClinicalTrials.gov v2 API, run by the US National Library of Medicine. It's free, public, and includes the 450,000+ trials registered under US law.
What the AI does (and doesn't do)
When you ask for a plain-English explanation, we send the trial's official description and eligibility criteria to OpenAI's gpt-4o-mini model. The model rewrites the text for a layperson and extracts eligibility checkpoints. We cache each explanation so we only pay for it once per trial.
The model never makes diagnoses, never recommends treatment, and never claims a patient will or won't qualify. Its output is clearly labeled as AI-generated and should be confirmed with a clinician or the trial's listed contact.
How fresh is the data?
The search results and trial detail pages revalidate every hour. If a trial's status changes on ClinicalTrials.gov (e.g. recruiting → completed), ours catches up within an hour.