Last updated — February 2026
Open Methodology
How Atlas Rare works:
Where the trial data comes from
ClinicalTrials.gov, updated in real time via their public API. Every recruiting trial in the US. We don’t curate or filter based on sponsor relationships.
How we match trials to you
Four stages: we expand your condition to include every clinical synonym. We hard-filter by age, sex, and recruiting status. Then we extract eligibility criteria using AI and compare them to your profile, returning only trials where you meet inclusion criteria or where exclusion criteria are unclear enough to discuss with your doctor.
What the AI does and doesn’t do
The AI reads eligibility text and compares it to your profile. It doesn't make medical decisions or know your full medical history. It makes mistakes. Every match includes exactly why we matched it so you and your care team can evaluate it.
How Atlas Rare learns
Every time you search for a trial, prepare for an appointment, or report an outcome, Atlas Rare learns something specific to you. Your conditions, your history, and what has actually worked for you shapes every result you get after. When writing prior authorization letters, we pull what we have learned about your insurer: which citations worked, which policy language is relevant, what your specific denial pattern looks like. Everything goes into making Atlas Rare and your outcomes better and better.
How often it updates
Trial data: weekly. Your profile: whenever you update it. Memory of your past searches: continuously.
What Basic Mode is
If you turn on Basic Mode in Settings, Atlas Rare uses no AI memory. Every session starts fresh. You get the same trial matching and research tools with no learning layer.
Who to contact with questions
This page is updated whenever our methodology changes. If something here doesn't match what you're experiencing, email us — we want to know.