When you generate a note, your text is processed in memory to produce the output, then discarded the moment the response is returned. We never write your clinical text to disk, and we keep no server-side history of your notes.
Generated notes are saved only in your own browser, on your device. They never reach our servers, and clearing your browser clears them.
Before any text leaves our backend, a redaction pass strips common identifiers — Emirates IDs, phone numbers, emails, MRNs, labelled names and dates of birth. You should still enter only the minimum detail needed; the redactor is a safeguard, not a substitute for your judgment. This applies to typed and dictated text — photos you upload are processed as-is and are not auto-redacted, so remove patient identifiers before uploading.
The optional Record feature turns a consultation into an English summary for your note box. Audio is processed in memory only — never written to disk on our servers — and transcribed by a speech-to-text provider (ElevenLabs, US) that our backend instructs to delete its copy immediately after processing. Recording always starts with you confirming the patient's (or guardian's) informed consent, and because voice cannot be auto-redacted, the redaction pass applies to the transcript text before summarisation.
Generation runs on Anthropic's commercial API, which does not train on inputs or outputs and automatically deletes them within 30 days. We don't run ads, we don't sell data, and we use no third-party tracking.
All traffic is encrypted over HTTPS/TLS. Account data is encrypted at rest, and only the operator has administrative access, behind multi-factor authentication. Because no clinical content is stored, there is little for an incident to expose.
Everything ClearPass produces is a draft for you to review, edit and sign off before it goes into a record or to an insurer. ClearPass is a documentation aid — the treating clinician remains responsible for the final content.