Security

Tenant isolation is enforced centrally at the data layer, not by a filter on each query that someone might forget to write. Every read and write against a clinical record is scoped to the signed-in doctor before it reaches the database, and attempts to reach a record belonging to another clinic fail rather than return data.

Passwords are stored only as bcrypt hashes — we cannot read yours, and neither could anyone who obtained the database. Password-reset links are single-use, expire in an hour, and are themselves stored only as a hash. Resetting a password revokes any mobile or API sessions, on the assumption that a reset may follow a compromise.

Reception staff sign in with their own username. They can view and book appointments and nothing else — no patient files, no consultations, no prescriptions, no settings. Deleting a receptionist ends their access immediately.

Who opened or changed which patient record, and when — including anything your reception staff did — is recorded and visible to you under Settings → Activity log. Entries cannot be edited or deleted from within the app, and survive deletion of the record they describe; an audit trail that could be erased would not be one.

The log records that a record was accessed, never its clinical content.

All traffic is served over HTTPS. Patient documents and investigation reports are stored in the database alongside the records they belong to, served only to the owning doctor, and never made publicly addressable.

The database is backed up daily on a rotating schedule, with off-site copies so a single failed machine cannot take your records with it.

Every change runs through automated type-checking, tests and a security scan before it can be merged. Dependency advisories are reviewed on every pull request against an explicit allowlist, so a new vulnerability cannot pass silently, and secret scanning runs on every commit.

The optional assists send the symptom text you are working on to a model provider to generate repertory search phrases. The rubrics you are shown always come from our database, so a suggestion can never be a rubric that does not exist — and the model never sees your patient list or proposes a remedy. Leave the feature off and nothing is sent.

We do not claim HIPAA compliance. HIPAA is United States law and has no application to an Indian clinic; software that advertises it to Indian doctors is usually saying something it has not been audited for. We are built to India's Digital Personal Data Protection Act and Rules, and the privacy policy says exactly how.

We have not yet completed an independent security audit. When we do, it will be named on this page.

If you find a vulnerability, please tell us at support@sabkacare.com before disclosing it publicly. We will acknowledge within one working day, keep you updated, and credit you if you would like.