How we translate clinical trials

Every trial page on Stella is built from data published by the U.S. National Library of Medicine's ClinicalTrials.gov (CT.gov) registry — the same record the trial team and sponsor publish for regulators. The substantive facts about a trial (what's being tested, who can qualify, where it's run, which phase it's in, who sponsors it) all originate there. Each Stella trial page links back to the canonical CT.gov record so you can verify any specific detail against the source.

Stella uses Anthropic's Claude to translate the dense, technical CT.gov record into plain English. Specifically, the AI rewrites:

• The trial's formal title into a plain-English version
• The trial's description into a 2-3 sentence overview a non-expert can follow
• The eligibility criteria into bullet points ("you likely qualify if…" / "you likely don't qualify if…")
• What participation actually looks like — visits, duration, what's involved

The AI does not invent trial data, recommend whether a trial is right for you, or give medical advice. It cannot tell you whether you'll qualify for a specific trial — only the trial team can, through their own screening process. It cannot tell you whether a drug is safe, effective, or appropriate for your situation. It is a translation tool, not a clinician.

Each trial page shows two dates: when CT.gov last updated the source record, and when Stella last translated it. A weekly refresh job re-translates trials whose CT.gov record has changed since our last run. If you're looking at a page, the translation was generated from CT.gov data as of the date shown.

Before contacting a trial site to express interest, the trial team will conduct their own screening — which always supersedes anything on Stella. We recommend verifying with them directly:

• Whether you actually qualify (Stella's plain-English eligibility bullets are a summary, not a determination)
• What the schedule, locations, and time commitment look like for your specific situation
• Anything about the drug, intervention, or study procedures that affects your health decision

Trial pages on Stella are AI-translated, which means mistakes can happen — a misread eligibility line, a confused drug name, a phase translated wrong. If you spot one, the easiest thing is to compare the Stella page against the CT.gov source linked at the top of every trial page. You can also email eric@stellatrials.com and we'll fix it. Stella's goal is to make trials more accessible, not to substitute for the trial team or your doctor — treating the content as a starting point and verifying specifics with the people running the trial is the safest way to use Stella.