Hosted in India. One private account per hospital. Every view logged. The AI sees the transcript, never the patient. Your data, exportable any day.
The hospital is the data fiduciary; Vaidyah is its processor. Consent is recorded at registration, per purpose, and can be withdrawn. Patients can ask for a copy or deletion through their hospital.
AI output is a draft in a draft table. A record exists only when a doctor signs it with a PIN. Signed records are immutable; corrections are addenda.
Records, backups and consultation audio are stored in India, encrypted. The AI model is reached through an enterprise cloud service; no patient identifiers are in the request, and nothing is used to train the model.
Every record belongs to one hospital and the database itself enforces it. One hospital can never see another's patients, whatever screen or report is used.
| Front desk | Nurse | Doctor | Billing | |
|---|---|---|---|---|
| Register, queue | ● | ○ | ○ | ○ |
| Vitals | ○ | ● | ● | ○ |
| Clinical notes | — | view | ● | — |
| Sign | — | — | PIN | — |
| Payments, refunds | ● | — | — | ● |
Roles decide the screens and the actions: front desk, nurse, doctor, lab, pharmacy, billing, insurance, administrator. Doctors sign with a personal PIN. Sessions on shared computers end when left idle.
The transcript, the specialty and the note structure. Never the name, mobile, UHID, ABHA or address. The draft returns to the hospital's account and is marked as a draft until a doctor signs.
A full backup every day, encrypted and kept separately. Restoring from it is practised regularly, so a bad day at the hospital is never a lost record.
Storage, recording and messages are separate consents, taken by signature or OTP and shown on the record. Withdrawing one stops that use at once and is logged.
The administrator can export every table as CSV and every document as PDF, any day, with the signed notes' hashes. Leaving Vaidyah is a download, not a negotiation.
We show every credential we hold in the badges above and do not claim certifications we do not have.
A leading AI model, reached through an enterprise cloud service. Requests carry no patient identifiers and are not used to train the model.
It is encrypted, used to produce the draft, and deleted after the retention period the hospital sets once the note is signed.
Yes. Send your questionnaire to the address in the footer. We answer in writing and can walk your IT person through the architecture.
Write to the address in the footer. We acknowledge within one working day and keep you informed until it is fixed.
Send it to us. We answer in writing, in plain language, within one working day.
Get in touch